## Thinking Path
> - Paperclip manages AI agents that perform work.
> - The paid runner matrix verifies complete runner behavior with real
providers.
> - Each matrix job currently repeats work on GitHub-hosted runners.
> - Paperclip has an ephemeral AWS runner fleet for trusted workflows.
> - The paid workflow needs a reviewed and fail-closed route to that
fleet.
> - This pull request adds that route and keeps the existing hosted
runner as the disabled-state fallback.
> - The benefit is faster paid campaigns with the same actor,
environment, and secret boundaries.
## Linked Issues or Issue Description
**What happened?**
The Runner Full-Stack E2E workflow always uses `ubuntu-latest-m`. It
limits the matrix to 57 parallel jobs. The repository AWS fleet can run
100 ephemeral jobs, but the paid workflow cannot select it.
**Expected behavior**
An explicit repository flag must select the reviewed AWS fleet label. A
missing or invalid flag must keep the existing hosted runner. The
workflow must authorize the stable actor identity before it routes any
paid job.
**Steps to reproduce**
1. Dispatch the Runner Full-Stack E2E workflow from `master`.
2. Inspect a paid matrix job.
3. Observe that the job requests `ubuntu-latest-m` even when the AWS
fleet should be used.
**Paperclip version or commit**
`da0947d3582ac7779d6bf11851c9938eca6c5c8c`
**Deployment mode**
GitHub Actions paid runner campaign.
## What Changed
- Add a fail-closed `RUNNER_E2E_AWS_ENABLED` switch.
- Select only the reviewed AWS fleet label or the existing hosted label.
- Permit up to 100 parallel jobs in AWS mode.
- Keep the hosted-runner limit at 57.
- Reauthorize paid execution before checkout and provider access.
- Stop paid checkouts from storing GitHub credentials.
- Cancel superseded validation-ref campaigns while preserving `master`
audit runs.
- Add workflow policy checks and operator documentation.
## Verification
- `git diff --check`
- `actionlint -ignore SC2129
.github/workflows/runner-full-stack-e2e.yml`
- The organization runner group permits this workflow only from
`refs/heads/master`.
- The repository AWS switch remains disabled until this pull request is
merged and a one-cell probe succeeds.
## Risks
- A wrong fleet policy can leave jobs queued. The disabled state keeps
the existing hosted runner.
- The AWS fleet uses paid compute. The workflow validates a configured
maximum of 100 jobs.
- The runner group, actor allowlist, and paid environment remain
separate enforcement layers.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
OpenAI Codex based on GPT-5 with agentic reasoning, repository
inspection, code editing, Git, GitHub API coordination, and static
workflow analysis. The exact deployed model identifier and
context-window size are not exposed to this task.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [ ] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the control plane for agents that perform work.
> - Paperclip Runner connects durable provider sessions to individual
task runs through PRP.
> - Provider continuity and per-run authority are different lifetimes.
> - The existing implementation mixed those lifetimes and lost event
metadata between provider frames, runnerd, persistence, API
sanitization, and the task thread.
> - That caused failed continuation, missing progress and Plans,
duplicate replies, hidden failures, and unsafe recovery.
> - This repair gives every heartbeat fresh authority, preserves
qualified provider-session continuity, and restores one lossless
presentation path without changing direct adapters.
## Linked Issues or Issue Description
**What happened?**
A second native heartbeat could reuse tickets, leases, command receipts,
sequence state, and run identity from the first heartbeat. Provider
phase and item identity could be lost before the UI read them. Redaction
could corrupt protocol discriminators while still missing malformed
credential tails. The task thread could fold progress into the final
response, hide failures, or show more than one final answer. Native
Codex also exposed approval modes that do not yet have a durable
approval bridge.
**Expected behavior**
Each heartbeat uses a new PRP authority epoch. Codex and OpenCode
preserve exact qualified provider sessions; ACPX emits an explicit
continuity event when its qualified process-replacement policy is used.
Every accepted provider event is presented, classified as internal, or
surfaced as unsupported. The task page shows chronological progress,
reasoning summaries, activity, Plans, interactions, terminal failures,
and exactly one final reply. Direct adapters retain their existing path.
**Steps to reproduce**
1. Enable the unified experimental Paperclip Runner setting.
2. Create a local native Codex, OpenCode, ACPX Claude, or ACPX Codex
agent.
3. Run response, Plan, structured-question/resume, restart,
cancellation, and failure scenarios.
4. Reload the task while active, waiting, failed, and settled.
5. On the old implementation, observe stale run authority, missing
classifications, incomplete output, or duplicated/folded replies.
**Paperclip version or commit**
The repair is based directly on `master` at
`87d05e194b643810d16d20612115acd01d735d43`.
**Deployment mode**
Local development with the embedded database.
Related work: Refs #12616, #12646, #12666, #12685, and #12700.
## What Changed
- Rotates PRP control-plane, outbox, ticket, lease, command, receipt,
and sequence authority for each heartbeat while carrying forward only a
validated provider-session identity.
- Reads `control-plane-state.json`, validates both durable schemas and
lifecycle values, resumes coherent current runs, archives qualified
settled authority, and quarantines malformed or mismatched scoped state
without moving ambiguous live legacy state.
- Preserves Codex provider phase and stable item identities so
commentary remains progress and only `final_answer` becomes final.
- Adds raw OpenCode HTTP/SSE boundary coverage and canonical reasoning
lifecycle mapping.
- Makes ACPX normalization lossless for visible reasoning, tool
lifecycle metadata, stable bounded identities, Plan revisions,
structured requests, failures, and qualified process replacement. Only
the compatible terminal assistant message is promoted as final.
- Applies schema-aware redaction before generic JWT-shaped detection and
scans every diagnostic string leaf. Malformed raw/escaped quoted
credential tails are redacted in both server and durable Rust state.
- Restores snapshot-style chronological task presentation, expandable
tool activity, inline Plan cards, visible waiting/resume/cancel/failure
states, and exactly one final answer.
- Makes `never` the only qualified native Codex permission mode and
rejects unsupported persisted native modes with remediation. OpenCode
and ACPX policies remain intact.
- Keeps the unified experimental Runner setting as the only enablement
flag. Onboarding and direct Codex, Claude, and OpenCode stay on their
legacy execution/finalization paths.
- Adds cross-language goldens, authority/recovery/fault coverage, exact
response/count assertions, and native plus legacy acceptance scenarios.
## Verification
- Pull-request GitHub Actions run Rust formatting/tests, TypeScript
checks, server/UI tests, builds, protocol drift checks, browser E2E, and
security scans.
- A separate workflow-only validation ref is pinned directly on this PR
head and runs the 35-cell paid local matrix: three core scenarios plus
structured-question resume and restart/resume for native Codex, native
OpenCode, ACPX Claude, ACPX Codex, and direct Codex/Claude/OpenCode.
Run: https://github.com/paperclipai/paperclip/actions/runs/33682434315
- Acceptance requires exact single visible replies, monotonic sequences,
matching envelope discriminators, one semantic terminal, one run
terminal, no unresolved interaction, no duplicate mutation, no secret
leakage, provider continuity, and zero native rows for direct adapters.
- Per maintainer direction, tests are running in GitHub Actions rather
than on the slower local host. Only formatters and static diff checks
were run locally.
## Risks
- Recovery from old or partial filesystem state is sensitive. The repair
fails closed, preserves active or unverifiable authority, and
quarantines only state whose scoped ownership is safe to move.
- Provider event formats can change. Closed validators and boundary
goldens turn new or malformed events into visible diagnostics instead of
silent drops.
- Shared task presentation could affect direct adapters. Runtime-fact
gating plus the direct-adapter matrix protect the existing path.
- Managed and remote providers are not qualified here. Shared code
continues to compile and fail safely, but live qualification is
deferred.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
OpenAI Codex based on GPT-5. The exact deployed snapshot and
context-window size are not exposed to this task. It used agentic
reasoning, repository inspection, code editing, Git, parallel subagents,
and GitHub Actions.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change and contains no internal
Paperclip ticket id
- [ ] I have run tests locally and they pass (intentionally deferred to
GitHub Actions)
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented risks above
- [ ] All Paperclip CI gates are green
- [ ] The paid local-provider matrix is green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The runner subsystem executes agent work across local and managed
provider backends.
> - The lower pull requests restore the task runtime, provider backends,
and managed-provider control plane.
> - The restored system needs repeatable full-stack checks before it can
ship safely.
> - Paid live checks also need clear access, cost, and secret controls.
> - This pull request adds acceptance, live evaluation, chaos, and
release gates for the restored runner stack.
> - The benefit is measurable runner parity with safer release
decisions.
## Linked Issues or Issue Description
**Subsystem affected**
Cross-cutting. This change covers runner tests, release workflows,
server contracts, and evaluation tools.
**Problem or motivation**
The runner stack did not have one complete acceptance surface for native
Codex, ACPX, Claude Managed, and AWS AgentCore. Release checks could
miss provider drift, task-view regressions, cost-policy errors, and
destructive cleanup errors.
**Proposed solution**
Add a 57-cell full-stack catalog, a Daytona image, and opt-in paid
workflows. Add live evaluation, chaos, cost-limit, redaction, and
release contract checks. Add AWS AgentCore infrastructure and guarded
provisioning tools. Keep the native runner experimental flag off by
default.
**Alternatives considered**
We considered manual smoke tests only. They do not give repeatable
evidence and they do not protect release branches. We also considered
one large pull request. The stacked pull requests keep each review below
the Greptile file limit.
**Roadmap alignment**
This work supports the shipped Cloud / Sandbox agents milestone and the
shipped Agent evals & feedback milestone in `ROADMAP.md`.
Related stack:
- #12699 adds managed provider backends and lifecycle support.
- #12691 adds qualified OpenCode and ACPX provider backends.
- #12685 restores task runtime rendering and steering.
## What Changed
- Add the runner full-stack harness with 57 catalog cells and 60 unit
tests.
- Add a Daytona runner image with digest-pinned base images and
base-aware image-content checks.
- Add guarded live evaluation and chaos workflows with a fixed
40-execution matrix; live and full-stack paid schedules now run only on
Sundays or by manual dispatch.
- Add in-flight reported-usage cost stops, post-turn cost caps,
exact-threshold failure classification, secret redaction, retry
classification, and actor authorization.
- Reattach stream and hard-budget listeners before restart-recovery
continuations so restored paid sessions cannot bypass in-flight
interruption.
- Preserve OpenCode usage and cost across tool-loop messages and turns
while exposing an explicit current-run delta to durable accounting.
- Keep PNG/WebM evidence in access-controlled artifacts only, reject
SVG, and publish only pruned inert structured per-attempt evidence.
- Add AWS AgentCore infrastructure, provisioning checks, and smoke
tools; reject unsafe model identifiers, require exact stack ownership
markers, and make failed-stack replacement explicit.
- Add evaluation-session contracts and capability reports.
- Add release workflow checks for immutable action pins, frozen
dependency installs, exact weekly cron shape, paid-run guards,
provider-secret isolation, and chaos test paths.
- Reauthorize the original and triggering numeric actor IDs as the first
step of every provider-secret job, including partial reruns, before
checkout or provider access.
- Give each full-stack matrix cell only its matching provider
credential, expose Daytona only to Daytona cells, and disable shared
dependency caches anywhere paid credentials or OIDC write access are
present.
- Protect the legacy manual E2E workflow with the same default-branch,
allowlist, environment, and per-job authorization boundary.
- Rotate live-eval candidates by week and retain 120 days of compatible
history so the seven-week trend window remains viable.
- Restore the root runner-acceptance commands and reconcile reported
snapshots,
raw receipts, and terminal usage without double counting or losing late
usage.
- Mark ACPX token deltas exact only when every budget field is present,
keep
cumulative cost/request authority separate, reject non-USD cost
labeling,
and include thought tokens in output-token budgets.
- Keep `enableNativeRunner` off by default. The acceptance harness
enables it only in its isolated test instance.
## Verification
Passed locally:
- `pnpm --filter @paperclipai/paperclip-runner typecheck`
- `pnpm test:runner-acceptance:typecheck`
- `pnpm test:runner-acceptance` (19 tests)
- focused OpenCode proxy, driver, runnerd transport, live-session, and
turn-stream tests (106 tests)
- `pnpm --filter @paperclipai/paperclip-runner exec vitest run
src/live/clean-room-server.test.ts` (22 tests)
- `pnpm test:e2e:runner:typecheck`
- `pnpm test:e2e:runner:unit` (62 tests)
- `node --test scripts/__tests__/release-verify-workflow.test.mjs`
- `pnpm --filter @paperclipai/paperclip-runner
test:runner-workflow-evals` (22 tests)
- `pnpm -r typecheck`
- `pnpm build`
- `node --test
packages/paperclip-runner/scripts/aws-agentcore-provisioning.test.mjs`
(6 tests)
- `git diff --check`
- `cargo test --manifest-path
packages/paperclip-runner/runner/Cargo.toml -p paperclip-runner-core
--lib --locked` (161 tests)
- focused ACPX provider-event tests (10 tests)
- The rebased PR changes 92 files. `pnpm-lock.yaml` is unchanged.
I did not run paid live provider jobs or provision AWS resources. Those
checks need credentials and can create cost.
## Risks
The paid workflows can create provider cost. They require an allowlisted
original and triggering actor, the protected `runner-e2e-paid`
environment, explicit opt-in variables, and cost limits. The four
provider credentials exist only in that master-only environment, which
requires allowlisted reviewer approval and disables administrator
bypass; repository and organization Actions scopes contain no copies.
Provider usage arrives after a billable request, so the live guard
cannot prevent one request from crossing a threshold. It interrupts
immediately on the first reported threshold hit and permits no
continuation.
Visual evidence can contain secrets rendered as pixels. PNG/WebM remain
only in access-controlled workflow artifacts; SVG and per-attempt XML
are excluded, and S3/Pages receive a pruned structured dashboard.
The AWS scripts can create cloud resources. They use explicit commands,
least-privilege roles, KMS encryption, saved nonsecret metadata, and
explicit teardown.
This pull request does not enable the experimental native runner for
existing instances.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
OpenAI Codex with GPT-5. The model used extended reasoning, tool use,
code execution, and parallel subagents.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The onboarding command starts the local server and opens the
first-run wizard.
> - Interactive onboarding stopped opening the browser by default.
> - Organization creation could also succeed in the API while the wizard
stayed on the name step.
> - The npm canary workflow did not prove that the published package
could complete this path.
> - This pull request repairs the startup and organization transitions
and adds an exact-version canary smoke gate.
> - The benefit is a quickstart that works for users and is tested after
each canary publish.
## Linked Issues or Issue Description
Related: #12557 covers a separate final-route onboarding handoff.
**What happened?**
Interactive `paperclipai onboard` runs did not open the onboarding page.
The organization API request could succeed while a same-company context
update caused the wizard to stay on the organization step. The canary
release lane did not test the exact published npm package through this
path.
**Expected behavior**
Interactive onboarding must open the browser once. A successful
organization request must advance to the first-agent step when the
surrounding context adopts the same organization. Each published canary
must install in a clean environment and reach the model connection step.
**Steps to reproduce**
1. Run `npx paperclipai@canary onboard --data-dir "$(mktemp -d
/tmp/paperclip-canary.XXXXXX)"` in an interactive terminal.
2. Enter an organization name while the company context refreshes from
the create response.
3. Observe that the browser does not open or that the wizard can remain
on the organization step after the API creates it.
4. Inspect the canary release lane and observe that no post-publish
onboarding test runs against the exact npm version.
**Paperclip version or commit**
The issue reproduced with `2026.901.0-canary.8` and the source state
before this pull request.
**Deployment mode**
Local trusted quickstart with embedded PostgreSQL. The install source
can be npm or a source checkout.
## What Changed
- Open the browser once for interactive foreground onboarding.
- Preserve explicit browser opt-outs and restore the prior environment
value after startup.
- Accept a same-company context update after organization creation and
reject a different-company takeover with an explicit error.
- Export the exact canary version from the publish job.
- Install and test that exact npm version in a clean Playwright smoke
job through the "Connect a model" step.
- Upload server logs, traces, screenshots, and the Playwright report
when the canary smoke fails.
- Document the interactive default and headless opt-outs.
## Verification
- `pnpm exec vitest run cli/src/__tests__/onboard.test.ts
ui/src/components/OnboardingWizard.step.test.tsx --reporter=dot` passes
with 37 tests.
- `node --test scripts/__tests__/release-verify-workflow.test.mjs`
passes with 9 tests.
- `pnpm exec playwright test --config tests/e2e/playwright.config.ts
tests/e2e/onboarding.spec.ts` passes with 2 tests.
- `PAPERCLIPAI_VERSION=2026.901.0-canary.8 pnpm run
test:canary-onboarding-smoke` passes against the published npm package.
- `pnpm check:token-gates` passes.
- `pnpm -r typecheck` passes.
- `pnpm build` passes.
- A fresh interactive source run opens the browser and reaches "Connect
a model" after organization and agent naming.
## Risks
- Low risk. Automatic browser opening only applies to interactive
foreground onboarding.
- `PAPERCLIP_NO_BROWSER=1` and `PAPERCLIP_OPEN_ON_LISTEN=false` keep
headless runs silent.
- A different organization context still blocks the pending create
transition.
- The canary package is immutable before the smoke runs. A smoke failure
leaves the package published but makes the release workflow red.
- This change does not modify REST APIs, database schemas, or shared
data types.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
- OpenAI Codex based on GPT-5. The runtime does not expose the exact
deployment snapshot or context-window size. The model used reasoning,
browser automation, repository tools, shell commands, code editing, and
test execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip uses GitHub Actions to verify pull requests before merge.
> - The pull request workflow calls a trusted reusable workflow at an
immutable commit SHA.
> - Pull request #12507 added stack-aware CI scope to that reusable
workflow.
> - The caller still points to the prior workflow commit, so the new
behavior is not active.
> - This pull request advances the caller to the merged commit from pull
request #12507.
> - The benefit is that native stack middle layers stop starting
redundant full CI matrices.
## Linked Issues or Issue Description
Refs #12507
**What existing behavior does this improve?**
The pull request workflow still calls the trusted CI definition that
predates stack-aware scope selection.
**Subsystem affected**
GitHub Actions pull request verification.
**Current behavior**
Every pull request layer in a native stack starts the complete test,
build, canary, and E2E matrix.
**Proposed behavior**
Call the trusted workflow from merged master commit
`39b8ee2960541d14b380f95365deecba6723d9bd`. That workflow runs full CI
for ordinary, top, and lowest-unmerged pull requests. It keeps policy
and required aggregate checks on middle layers.
**Reason and benefit**
This completes the two-step immutable workflow rollout from pull request
#12507. Large stacks will use fewer runners and will spend less time
waiting for duplicate jobs.
**Breaking changes**
Middle native stack layers no longer run the complete CI matrix.
Required aggregate checks remain present and fail closed if the stack
scope is missing or invalid.
## What Changed
- Pin `.github/workflows/pr.yml` to merged master commit
`39b8ee2960541d14b380f95365deecba6723d9bd`.
- Activate the stack-aware trusted workflow that merged in pull request
#12507.
## Verification
- `node --test scripts/__tests__/e2e-shard.test.mjs` — 11 tests passed.
- `actionlint .github/workflows/pr.yml .github/workflows/pr-trusted.yml`
— passed.
- `git diff --check origin/master...HEAD` — passed.
- Confirm that the caller SHA equals the merge commit for pull request
#12507.
## Risks
- The caller is immutable and points to a commit that exists on
`master`.
- Ordinary pull requests and merge-relevant stack layers still run full
CI.
- A rollback can restore the prior immutable SHA in one line.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
- OpenAI Codex with GPT-5. The exact deployment suffix and context
window are not exposed. The model used reasoning, repository tools, code
execution, Git, and GitHub API access.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip uses GitHub Actions to protect changes before they enter
`master`.
> - GitHub evaluates every pull request in a native stack against the
stack base.
> - The current workflow therefore starts the complete CI matrix for
every layer in a stack.
> - A large stack can queue many copies of the same integrated
verification and delay every pull request.
> - GitHub provides stack position and base metadata so workflows can
select merge-relevant layers.
> - This pull request keeps policy and required check names on every
layer, but runs full CI only for ordinary pull requests, the top layer,
and the lowest unmerged layer.
> - The benefit is much lower CI load without weakening the
required-check contract.
## Linked Issues or Issue Description
**What existing behavior does this improve?**
The trusted pull request workflow currently runs every test, build,
canary, and E2E lane for every pull request in a native stack.
**Subsystem affected**
GitHub Actions pull request verification.
**Current behavior**
A stack with 61 pull requests can start 61 complete CI matrices after a
cascading rebase.
**Proposed behavior**
Run the always-on policy job and stable required-check aggregators for
every layer. Run the complete verification matrix only for ordinary pull
requests, the top stack layer, and the lowest unmerged stack layer.
**Reason and benefit**
The top layer verifies the integrated stack. The lowest unmerged layer
verifies the current merge candidate. Middle layers keep
branch-protection checks without consuming the complete runner matrix.
**Breaking changes**
Middle stack layers no longer run the complete CI matrix. Their `ci /
verify` and `ci / e2e` checks still require the policy job to pass and
require every expensive lane to be intentionally skipped.
## What Changed
- Add a fail-safe stack scope decision to the trusted PR runner gate.
- Run typecheck, general tests, build, serialized tests, canary, and E2E
shards only for ordinary, top, and lowest-unmerged pull requests.
- Preserve the required `ci / verify` and `ci / e2e` names on every
layer.
- Make the required aggregators distinguish valid middle-layer skips
from failures or missing scope decisions.
- Add regression coverage for ordinary, top, bottom, middle, and
malformed stack metadata.
## Verification
- `node --test scripts/__tests__/e2e-shard.test.mjs` — 11 tests passed.
- `actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml`
— passed.
- `git diff --check origin/master...HEAD` — passed.
- The caller remains pinned to the current trusted workflow. A separate
activation change must advance the immutable SHA after this pull request
lands.
## Risks
- Incorrect stack classification could skip important jobs. Missing or
malformed stack metadata defaults to full CI.
- Middle-layer required checks depend on the policy job and verify that
all expensive jobs have the `skipped` result.
- The reusable workflow change does not become active until the
immutable caller SHA advances in a separate change.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
- OpenAI Codex with GPT-5. The exact deployment suffix and context
window are not exposed. The model used reasoning, repository tools, code
execution, Git, and GitHub API access.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` /
`Closes: #` / `Refs: #` OR (b) described the issue in-PR following the
relevant issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip uses trusted GitHub Actions workflows to verify each pull
request.
> - The caller pins the reusable workflow to an immutable merged commit.
> - Native stacked pull requests can inherit a stale lockfile from a
parent layer.
> - Pull request #12461 added safe merge-tree lockfile regeneration to
the trusted workflow.
> - The caller must now select that merged workflow version.
> - This pull request advances the immutable pin to the merge commit
from #12461.
> - The benefit is reliable verification for native stacked pull
requests.
## Linked Issues or Issue Description
Refs #12461
## What Changed
- Pin the trusted pull request workflow to merged commit
`1da6b37fc56dacf5e7ffbd31da35756a1cba41f8`.
- Activate stale lockfile regeneration for native stacked pull requests.
## Verification
- `actionlint .github/workflows/pr.yml`
- `node --test scripts/__tests__/e2e-shard.test.mjs`
## Risks
- Low risk. This changes only the immutable reusable-workflow pin.
- The target commit is merged into `master` and passed all required
checks.
- The target workflow adds one lockfile-only install to the policy job.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
OpenAI Codex with model `gpt-5`, reasoning, tool use, and code
execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip uses trusted GitHub Actions workflows to verify every pull
request
> - Native stacked pull requests use another pull request branch as
their base
> - A parent layer can change a package manifest without committing
`pnpm-lock.yaml`
> - A child layer can inherit that manifest change without changing a
manifest itself
> - The current policy skips lockfile regeneration for that child and
downstream frozen installs fail
> - This pull request validates the complete merge tree and shares a
regenerated lockfile only when needed
> - The benefit is reliable stacked pull request verification without
weakening the trusted workflow boundary
## Linked Issues or Issue Description
**What happened?**
A stacked child pull request inherited a package manifest change from
its parent. The child did not change a manifest itself. The policy job
skipped lockfile regeneration. Downstream jobs tried to restore an
artifact that did not exist and then failed during frozen dependency
installation.
**Expected behavior**
The policy job must validate the complete pull request merge tree. It
must upload a regenerated lockfile when the checked-in lockfile is
stale, including on a stacked child layer.
**Steps to reproduce**
1. Create a parent pull request that changes `package.json` without
committing `pnpm-lock.yaml`.
2. Create a child pull request on that branch without another manifest
change.
3. Run the trusted pull request workflow for the child.
4. Observe that frozen dependency installation fails because no
`pr-lockfile` artifact exists.
**Paperclip version or commit**
`f173ee09fa5c2ced7806bba47b54c3df853ab4df`
**Deployment mode**
GitHub Actions trusted pull request workflow.
**Agent adapter(s) involved**
Not adapter-specific. This is a core CI workflow bug.
## What Changed
- Regenerate the lockfile from every checked-out merge tree.
- Compare the generated lockfile with the checked-in copy before upload.
- Download the artifact only when the policy job reports that it
uploaded one.
- Fail closed when a reported artifact is missing.
- Add a workflow contract test for stacked lockfile handling.
- Keep the caller pinned to the last merged trusted SHA; after this
implementation merges, a separate activation PR will advance the
immutable pin to its merge commit.
## Verification
- `actionlint .github/workflows/pr-trusted.yml`
- `node --test scripts/__tests__/e2e-shard.test.mjs`
## Risks
- The policy job runs one lockfile-only install for every pull request.
This can add a small amount of CI time.
- A missing artifact now fails immediately when the policy job reports
an upload. This is intentional because it exposes workflow corruption.
- No runtime or product behavior changes.
- The implementation/activation split is intentional: unmerged
PR-authored workflow code must never execute on trusted runners.
## Model Used
OpenAI Codex with model `gpt-5`, reasoning, tool use, and code
execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip uses pull request checks to protect changes.
> - The trusted workflow selects GitHub or AWS runners.
> - Pull request #12462 removed an unreliable stale API equality.
> - The public caller must pin an immutable trusted workflow commit.
> - This pull request changes only that pin.
> - The benefit is correct AWS routing for trusted stacked pull
requests.
## Linked Issues or Issue Description
Refs #12339
Refs #12462
## What Changed
- Pin the thin pull request caller to
f9c32513b2.
## Verification
- actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml
- github-runners/tests/test-workflow.sh .github/workflows/pr-trusted.yml
.github/workflows/pr.yml
- The runner group authorizes this SHA and all in-flight prior SHAs.
## Risks
- Low risk. This changes one immutable workflow reference.
- Older workflow SHAs remain authorized while queued and active jobs
finish.
## Model Used
- OpenAI Codex, GPT-5.6, with reasoning and terminal tool use.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used with version and capability
details
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have linked existing public pull requests
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
ticket ID
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation
- [x] I have considered and documented risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open findings
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip uses pull request checks to protect changes.
> - The trusted CI gate selects GitHub or AWS runners.
> - GitHub can return an old pull request base SHA after the live base
advances.
> - The signed event, live Git ref, ancestry, and merge parents provide
the required proof.
> - The stale API field rejects a safe run even when those proofs pass.
> - This pull request removes that unreliable equality.
> - The benefit is correct AWS routing for trusted stacked pull
requests.
## Linked Issues or Issue Description
Refs #12339
Refs #12459
## What Changed
- Stop treating pull request base.sha as a current-state signal.
- Keep the signed event base SHA and live ref descendant check.
- Keep the live base or synthetic base merge-parent proof.
- Keep all numeric identity, repository, head SHA, and triggering actor
checks.
## Verification
- actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml
- github-runners/tests/test-workflow.sh .github/workflows/pr-trusted.yml
.github/workflows/pr.yml
- The corrected gate selected the Fleet label with the exact live event
data from PR #12339 run 33201610330.
## Risks
- The pull request API base SHA can be stale and is no longer compared.
- Replaced ancestry, a changed head, a changed merge parent, and a
changed merge tree still fail closed.
## Model Used
- OpenAI Codex, GPT-5.6, with reasoning and terminal tool use.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used with version and capability
details
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have linked existing public pull requests
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
ticket ID
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation in the operations repository
- [x] I have considered and documented risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open findings
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip uses pull request checks to protect changes.
> - The trusted CI workflow selects GitHub or AWS runners.
> - Pull request #12459 fixed validation when a stacked base advances.
> - The public caller must use an immutable trusted workflow SHA.
> - This pull request changes only that SHA.
> - The benefit is automatic AWS routing for the affected trusted stack
runs.
## Linked Issues or Issue Description
Refs #12339
Refs #12459
## What Changed
- Pin the thin pull request caller to
f929355fb9.
## Verification
- actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml
- github-runners/tests/test-workflow.sh .github/workflows/pr-trusted.yml
.github/workflows/pr.yml
- The runner group already authorizes the old and new exact workflow
SHAs.
## Risks
- Low risk. This changes one immutable workflow reference.
- The old workflow SHA remains authorized while current runs finish.
## Model Used
- OpenAI Codex, GPT-5.6, with reasoning and terminal tool use.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used with version and capability
details
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have linked existing public pull requests
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
ticket ID
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation
- [x] I have considered and documented risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open findings
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip uses pull request checks to protect changes.
> - The trusted CI workflow selects GitHub or AWS runners.
> - Stacked pull requests can advance their base branch while a gate
waits.
> - The gate already proves that the live base descends from the event
base.
> - An earlier exact base check rejects that safe state before the
ancestry check runs.
> - This pull request removes the conflicting check and verifies live
API consistency.
> - The benefit is automatic AWS routing for trusted stacked pull
requests without weaker identity checks.
## Linked Issues or Issue Description
Refs #12339
Refs #12457
## What Changed
- Allow the live pull request base SHA to advance from the signed event
base snapshot.
- Require the pull request API base SHA to match the live Git ref during
validation.
- Keep the numeric author, sender, and triggering actor checks.
- Keep descendant ancestry and synthetic merge validation.
## Verification
- actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml
- github-runners/tests/test-workflow.sh .github/workflows/pr-trusted.yml
.github/workflows/pr.yml
- The routing suite covers a live stacked base advance and a base change
during validation.
## Risks
- A trusted stacked run can use a newer descendant base than its signed
event snapshot.
- Replaced ancestry still fails closed.
- A live base change during gate validation still fails closed.
## Model Used
- OpenAI Codex, GPT-5.6, with reasoning and terminal tool use.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either linked existing issues or described the issue in
this pull request
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
ticket ID
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation in the operations repository
- [x] I have considered and documented risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open findings
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Pull request checks protect the application and its contributors.
> - The trusted CI gate must validate both direct and stacked GitHub
merge shapes.
> - Pull request #12457 added that validation at an immutable master
SHA.
> - The active caller still pins the prior workflow version.
> - This pull request pins the caller to the newly authorized SHA.
> - The benefit is safe automatic AWS routing for trusted stacked pull
requests.
## Linked Issues or Issue Description
Refs #12457
**What existing behavior does this improve?**
The active caller uses a gate that fails closed on GitHub synthetic
stacked merge parents.
**Subsystem affected**
GitHub Actions pull request routing.
**Current behavior**
Trusted stacked pull requests run on GitHub-hosted runners after the
merge-parent check rejects the synthetic base merge.
**Proposed behavior**
The caller uses the authorized workflow SHA
7b199fcafa.
**Reason and benefit**
The change lets approved stack jobs use the 100-runner AWS Fleet while
all numeric identity and live-state checks remain active.
**Breaking changes**
None for untrusted contributors. Their jobs continue on GitHub-hosted
runners.
**Additional context**
The runner group authorizes both this SHA and the previous SHA during
rotation.
## What Changed
- Pin the thin caller to the authorized stacked-merge validator SHA.
## Verification
- Ran actionlint on both pull request workflows.
- Ran the external routing suite against the trusted workflow and
caller.
- Verified that the runner group authorizes both immutable workflow
versions.
## Risks
A bad pin can stop CI. The target is on master, the previous SHA stays
authorized, and the caller change is a one-line reversible pin update.
> For core feature work, check ROADMAP.md first and discuss it in #dev
before opening the PR. Feature PRs that overlap with planned core work
may need to be redirected — check the roadmap first. See
CONTRIBUTING.md.
## Model Used
OpenAI Codex on GPT-5. The exact deployment ID and context-window size
are not exposed. The model used reasoning, tool use, GitHub API access,
and local code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used with version and capability
details
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have linked an existing public item and described the issue
in-PR
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
Paperclip ticket id
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Pull request checks protect the application and its contributors.
> - The trusted CI gate verifies signed event data against live GitHub
state.
> - GitHub gives a stacked pull request a synthetic merge commit for its
base stack.
> - The current validator expects the raw base SHA as the first merge
parent.
> - That assumption sends safe stacked pull requests to GitHub-hosted
runners.
> - This pull request validates the synthetic base merge and its
ancestry explicitly.
> - The benefit is safe automatic AWS routing for approved stacked pull
requests.
## Linked Issues or Issue Description
Refs #12455
Refs #12456
**What existing behavior does this improve?**
The trusted runner gate validates merge commits for master-based pull
requests but rejects GitHub synthetic base merges for stacked pull
requests.
**Subsystem affected**
GitHub Actions pull request identity and merge-state validation.
**Current behavior**
A trusted stacked pull request passes all numeric identity checks. The
gate fails closed because the first event merge parent is a GitHub
synthetic base merge instead of the raw base snapshot SHA.
**Proposed behavior**
The gate verifies the current base-ref tip, base-snapshot ancestry,
identical event and live merge parents, the child head parent, the
synthetic base merge parents, and identical event and live merge trees.
**Reason and benefit**
The change preserves fail-closed live-state validation while allowing
approved stacked pull requests to use the isolated AWS Fleet.
**Breaking changes**
None for untrusted contributors. Trusted stacked pull requests can
select AWS after the caller pins this workflow version.
**Additional context**
GitHub builds a stacked test merge in two steps. It first merges the
stack base into its own current base. It then uses that synthetic commit
as the first parent of the child test merge.
## What Changed
- Fetch and validate the current base branch ref.
- Require the event base snapshot to remain an ancestor of that ref.
- Validate direct and synthetic base merge parent shapes.
- Preserve the existing child-head, live-state, merge-tree, repository,
and actor checks.
## Verification
- Ran actionlint on the trusted workflow.
- Ran the external routing suite.
- Added positive coverage for the GitHub stacked merge shape.
- Added fail-closed coverage for replaced base ancestry and a synthetic
merge that omits the current base ref.
- Replayed the checks against the live merge shape for pull request
#12340.
## Risks
The gate has more GitHub API reads. Its five-minute timeout and
fail-closed behavior limit the effect of API errors. The accepted
synthetic commit must be the same in the event and live merge, must
include the current base branch as a direct parent, and must produce the
same merge tree.
> For core feature work, check ROADMAP.md first and discuss it in #dev
before opening the PR. Feature PRs that overlap with planned core work
may need to be redirected — check the roadmap first. See
CONTRIBUTING.md.
## Model Used
OpenAI Codex on GPT-5. The exact deployment ID and context-window size
are not exposed. The model used reasoning, tool use, GitHub API access,
and local code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used with version and capability
details
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have linked existing public items and described the issue in-PR
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
Paperclip ticket id
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Pull request checks protect the application and its contributors.
> - The trusted CI workflow sends approved contributors to isolated AWS
runners.
> - Pull request #12455 added safe support for non-master base branches.
> - The active caller still filters for master and uses the previous
immutable workflow SHA.
> - This pull request removes the base-branch trigger filter and pins
the caller to the authorized stack-aware SHA.
> - The benefit is that trusted stacked pull requests can use the AWS
fleet automatically.
## Linked Issues or Issue Description
Refs #12455
**What existing behavior does this improve?**
The active pull request caller only triggers for master and selects the
master-only trusted workflow version.
**Subsystem affected**
GitHub Actions pull request routing.
**Current behavior**
Trusted stacked pull requests either do not trigger the default caller
or use a branch-local older caller. Their heavy jobs remain in the
GitHub-hosted queue.
**Proposed behavior**
The caller triggers for every pull request base branch. It uses the
authorized stack-aware workflow at full SHA
d6b33d6c16.
**Reason and benefit**
The change sends heavy jobs from approved stacked pull requests to the
100-runner AWS Fleet. Unlisted contributors still receive the normal
GitHub-hosted checks.
**Breaking changes**
Trusted pull requests with non-master base branches now use AWS.
Untrusted pull requests continue to use GitHub-hosted runners.
**Additional context**
The runner group currently authorizes both the old and new immutable
workflow SHAs. The old SHA stays authorized until outstanding runs
finish.
## What Changed
- Remove the caller base-branch filter so stacked pull requests trigger
CI.
- Pin the thin caller to the authorized stack-aware reusable workflow
SHA.
## Verification
- Ran actionlint on both pull request workflow files.
- Ran the external routing test suite against the new reusable workflow
and thin caller.
- Verified that the runner group authorizes both immutable SHAs and only
the Paperclip repository.
## Risks
A bad pin can stop pull request CI. The target SHA is on master,
contains the reviewed reusable workflow, and is already authorized
beside the prior SHA. This gives the rotation a rollback path. Every
untrusted run still uses a GitHub-hosted gate and route.
> For core feature work, check ROADMAP.md first and discuss it in #dev
before opening the PR. Feature PRs that overlap with planned core work
may need to be redirected — check the roadmap first. See
CONTRIBUTING.md.
## Model Used
OpenAI Codex on GPT-5. The exact deployment ID and context-window size
are not exposed. The model used reasoning, tool use, GitHub API access,
and local code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used with version and capability
details
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have linked an existing public item and described the issue
in-PR
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
Paperclip ticket id
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Pull request checks protect the application and its contributors.
> - The trusted CI workflow sends approved contributors to isolated AWS
runners.
> - The workflow currently limits AWS routing to pull requests that
target master.
> - Stacked pull requests target another branch and stay in the GitHub
queue.
> - This pull request keeps the identity checks and accepts a live
nonempty base branch in the Paperclip repository.
> - The benefit is that trusted stacked pull requests can use the AWS
fleet automatically.
## Linked Issues or Issue Description
**What existing behavior does this improve?**
The trusted pull request runner gate currently sends only master-based
pull requests to AWS.
**Subsystem affected**
GitHub Actions pull request routing.
**Current behavior**
A trusted contributor can pass all numeric identity checks. The gate
still selects GitHub-hosted runners when the pull request targets
another branch in a stack.
**Proposed behavior**
The gate accepts any nonempty base branch in the Paperclip base
repository. It still verifies the live base ref, base SHA, head SHA,
merge SHA, author ID, sender ID, and triggering actor ID.
**Reason and benefit**
Large pull request stacks currently add all heavy jobs to the limited
GitHub-hosted queue. This change lets approved contributors use the
isolated AWS fleet for those jobs.
**Breaking changes**
Trusted pull requests that target a non-master branch now use AWS
instead of GitHub-hosted runners. Untrusted pull requests keep the
current GitHub-hosted route.
**Additional context**
Pull request #12339 is the master-based first item in a stack. Its child
pull requests show this queue pattern.
## What Changed
- Replace the master-only route check with a nonempty base-ref check.
Keep the existing base-repository and live-state checks.
## Verification
- Ran actionlint on .github/workflows/pr-trusted.yml.
- Ran the external routing test suite. It passed the trusted
stacked-base case and all fail-closed identity cases.
## Risks
The AWS trust boundary now includes code from a non-master base branch
when the current pull request author, sender, and triggering actor are
all approved numeric GitHub IDs. An approved account can already submit
arbitrary head code. The runner remains isolated and has no production
access or repository secrets.
> For core feature work, check ROADMAP.md first and discuss it in #dev
before opening the PR. Feature PRs that overlap with planned core work
may need to be redirected — check the roadmap first. See
CONTRIBUTING.md.
## Model Used
OpenAI Codex on GPT-5. The exact deployment ID and context-window size
are not exposed. The model used reasoning, tool use, GitHub API access,
and local code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with Fixes: # / Closes #
/ Refs # OR (b) described the issue in-PR following the relevant issue
template
- [x] I have not referenced internal/instance-local Paperclip issues or
links
- [x] My branch name describes the change and contains no internal
Paperclip ticket id
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Summary
- validate the event base ref/SHA against the live PR state instead of
requiring the moving `master` branch tip to remain unchanged while a
hosted gate queues
- retain exact event/live merge parent and tree validation, plus
author/sender/rerun checks
## Canary finding
A seven-minute hosted-gate queue allowed `master` to advance. Requiring
the live branch tip to equal the event base snapshot would route
otherwise valid trusted runs back to GitHub-hosted indefinitely on a
busy repository.
## Validation
- actionlint and workflow-contract tests pass
- internal routing harness passes
- replaced PR base snapshot, stale head, changed merge parent/tree, and
untrusted actors all remain fail-closed
- AWS routing remains disabled during rotation
## Summary
- use `${{ github.sha }}` as the event merge commit validated by the
trusted gate
- retain exact base/head parent and identical-tree comparison against
the current live merge ref
## Canary finding
GitHub leaves `pull_request.merge_commit_sha` empty on some `opened`
payloads even though the workflow runs against a valid merge ref. The
gate safely fell back to GitHub-hosted and no EC2 instance launched.
## Validation
- `actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml`
- `node --test ./scripts/__tests__/e2e-shard.test.mjs`
- internal routing harness passes and asserts `EVENT_MERGE_SHA` is
sourced from `${{ github.sha }}`
- AWS routing remains disabled during rotation
## Summary
- rotate the thin PR caller from
`3b295b05dc8c8dd82c12e4a9c6f721446c5cb2e8` to
`b88fadb0390d2113933d5d155f16b07cbd5dafee`
- keep both exact SHAs authorized during the rotation
- keep AWS routing disabled until the caller and boundary verify
## Validation
- `actionlint .github/workflows/pr.yml .github/workflows/pr-trusted.yml`
- `node --test ./scripts/__tests__/e2e-shard.test.mjs`
- internal routing harness requires exactly one runner output per gate
execution
## Summary
- emit exactly one `runner` job output from the trusted gate
- write `ubuntu-latest` only inside fail-closed paths
- write the Fleet label only after every identity, PR-state,
merge-equivalence, and rerun-actor check passes
## Canary finding
The live gate reached the trusted success notice, but GitHub retained
the first of two duplicate `runner=` outputs, so policy still requested
`ubuntu-latest`. No EC2 instance launched. Routing was disabled
immediately.
## Validation
- `actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml`
- `node --test ./scripts/__tests__/e2e-shard.test.mjs`
- internal routing harness passes and now requires exactly one runner
output in all cases
- AWS routing remains disabled during rotation
## Summary
- rotate the thin PR caller from
`d9fc93d8383ece6fba721881a7aba638867f4996` to
`3b295b05dc8c8dd82c12e4a9c6f721446c5cb2e8`
- keep both exact workflow SHAs authorized in the runner group during
the rotation
- keep `AWS_CI_ENABLED=false` until this caller is merged and verified
## Validation
- `actionlint .github/workflows/pr.yml .github/workflows/pr-trusted.yml`
- `node --test ./scripts/__tests__/e2e-shard.test.mjs`
- full AWS/GitHub boundary verification passes with zero active runners
## Summary
- validate the live master ref and event base SHA before AWS routing
- accept GitHub synthetic merge commits only when the event and live
commits have the exact expected base/head parents and identical tree
- preserve fail-closed routing for malformed, stale, replaced, or
untrusted events
## Canary finding
A trusted reopened PR produced two synthetic merge SHAs with different
timestamps but identical current base/head parents and tree. The former
exact-SHA comparison safely fell back to GitHub-hosted runners, but
could not route a valid event to AWS.
## Validation
- `actionlint .github/workflows/pr-trusted.yml .github/workflows/pr.yml`
- `node --test ./scripts/__tests__/e2e-shard.test.mjs`
- real-event gate simulation selects the Fleet label for equivalent
merge commits
- negative simulations keep an untrusted sender and replaced head on
`ubuntu-latest`
- AWS routing remains disabled during rotation
## Thinking Path
> - Paperclip uses pull request CI to validate each proposed change
> - The existing workflow defines every heavy job in a PR-controlled
file
> - A trusted reusable workflow now contains the synchronized CI
definition
> - The caller must use an immutable default-branch SHA
> - This pull request replaces the duplicate job list with that pinned
caller
> - The benefit is automatic secure runner selection without workflow
drift
## Linked Issues or Issue Description
Refs #12436
Refs #12438
**What existing behavior does this improve?**
This improves how the pull request workflow selects trusted CI capacity.
**Subsystem affected**
Cross-cutting CI automation.
**Current behavior**
The active workflow contains a duplicate list of all heavy jobs. It
cannot use the administrator-controlled runner gate.
**Proposed behavior**
The active workflow calls the synchronized trusted workflow at an
immutable SHA. The trusted workflow selects GitHub-hosted or isolated
AWS capacity from the validated contributor identity.
**Reason and benefit**
The thin caller prevents pull request changes from replacing the
external-runner security gate. It also keeps runner selection automatic.
**Breaking changes**
The check names gain the reusable workflow job prefix. AWS routing
remains disabled until the canary starts.
## What Changed
- Replaced the duplicated heavy CI job list with one reusable-workflow
call.
- Pinned the call to the reviewed default-branch commit.
- Limited the caller token to actions, contents, and pull request read
access.
## Verification
- actionlint on both workflow files
- Trusted-routing tests
- Confirmed the pinned SHA contains the workflow and is an ancestor of
master
- Full AWS and GitHub runner-boundary verification with routing disabled
## Risks
The check context names change when GitHub expands the reusable
workflow. The rollout verifies the new aggregate contexts before branch
rules change. The repository kill switch remains off during this pull
request.
## Model Used
OpenAI Codex with GPT-5, tool use, and code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either linked a related public PR or described the issue
with the matching template fields
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
Paperclip ticket id
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [ ] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip uses pull request CI to protect changes before merge
> - The trusted reusable workflow will select isolated AWS capacity
> - The active workflow changed while the reusable workflow waited for
merge
> - The reusable policy must contain every current CI policy step before
activation
> - This pull request synchronizes the migration-order check and shell
validation
> - The benefit is one reviewed workflow version with verified job
parity
## Linked Issues or Issue Description
Refs #12436
**What existing behavior does this improve?**
This improves the pull request CI workflow synchronization before AWS
runner activation.
**Subsystem affected**
Cross-cutting CI automation.
**Current behavior**
The active workflow validates migration order. The new reusable workflow
does not yet contain that check.
**Proposed behavior**
Both workflow definitions contain the same heavy jobs and policy steps
before the active workflow becomes a thin caller.
**Reason and benefit**
The synchronization prevents policy drift during the two-step secure
rollout.
**Breaking changes**
None. AWS routing remains disabled.
## What Changed
- Added the current migration-order validation to the trusted workflow.
- Added the existing shellcheck intent annotation to the active
workflow.
- Verified normalized heavy-job parity between both definitions.
## Verification
- actionlint on both workflow files
- Local trusted-routing and normalized workflow-parity tests
- git diff --check
## Risks
Low risk. The migration check already runs in active CI. This change
copies it into the inactive trusted definition. AWS routing stays
disabled.
## Model Used
OpenAI Codex with GPT-5, tool use, and code execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either linked a related public PR or described the issue
with the matching template fields
- [x] I have not referenced internal or instance-local Paperclip issues
or links
- [x] My branch name describes the change and contains no internal
Paperclip ticket id
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [ ] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Pull request checks protect the quality of the control plane
> - The current checks depend only on the shared GitHub-hosted runner
limit
> - Busy periods leave many pull request jobs queued even when external
capacity is available
> - Public pull request code must not select or directly access private
runner infrastructure
> - This pull request adds an inactive reusable workflow with a
fail-closed identity gate
> - A later pull request can pin this workflow by its full master commit
SHA
> - The benefit is automatic, controlled access to isolated runner
capacity without changing current CI during bootstrap
## Linked Issues or Issue Description
**What existing behavior does this improve?**
This improves the pull request CI workflow. It prepares the existing
checks to use an administrator-controlled runner selection.
**Subsystem affected**
Cross-cutting GitHub Actions CI configuration.
**Current behavior**
Every pull request job uses `ubuntu-latest`. Jobs wait when the
GitHub-hosted concurrency limit is full.
**Proposed behavior**
Add a reusable copy of the current PR workflow. A GitHub-hosted gate
validates durable numeric user IDs and current GitHub API state. The
gate emits one runner label. The default and every validation failure
use `ubuntu-latest`. The AWS label is possible only when an
administrator enables it and every identity check passes.
This bootstrap pull request does not change the active
`.github/workflows/pr.yml` caller. A follow-up change will call this
workflow by the full master commit SHA.
**Reason and benefit**
The split bootstrap creates an immutable trust boundary before external
runners are reachable. It also keeps CI automatic for contributors.
Contributors do not select a runner.
**Breaking changes**
None in this bootstrap pull request. The active PR workflow does not
change.
## What Changed
- Added an inactive `workflow_call` copy of the current PR checks.
- Added a GitHub-hosted routing gate that checks the repository ID, pull
request author ID, event sender ID, rerun actor ID, base branch, head
SHA, merge SHA, and current pull request state.
- Made every validation failure select `ubuntu-latest`.
- Pinned every third-party action to a full commit SHA.
- Disabled persistent checkout credentials for all jobs.
- Limited the workflow token to Actions read, contents read, and pull
request read access.
## Verification
- `actionlint .github/workflows/pr-trusted.yml`
- Ran the dedicated workflow routing test harness against
`.github/workflows/pr-trusted.yml`.
- Compared the job keys with `.github/workflows/pr.yml`. The new
workflow contains every existing job plus the gate.
- Verified each pinned action commit against its current GitHub
major-version tag.
## Risks
The gate could route a trusted pull request to the wrong runner if an
identity check is incomplete. The gate checks durable numeric IDs from
the event and current GitHub API state. It checks the rerun actor
separately. It defaults to GitHub-hosted capacity before any validation
runs.
This file is inactive in this pull request. The follow-up caller and
runner-group restriction must use the exact commit that reaches
`master`.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
OpenAI Codex based on GPT-5. The exact serving model ID and
context-window size are not exposed in this environment. The model used
high-reasoning, terminal, GitHub API, browser, and web-research
capabilities.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - A release is gated by the release smoke: it installs the published
`paperclipai` artifact into a Docker container and drives the sign-in →
onboarding → first-agent path with Playwright
> - That suite runs only from the release pipeline, never on a pull
request, so it sees the UI only after the UI has already changed
> - The onboarding wizard was rebuilt into the agent arc. The "Name your
organization" step, the "Start Onboarding" launcher, and the agent role
picker are all gone
> - The spec still waited for those, so it failed on its first assertion
and blocked every nightly and beta release
> - The failure was also hard to read. The workflow uploaded no
container logs, because it learned the container's name only after the
harness succeeded, and the harness ran the container with `--rm` and
deleted it before anything read it
> - This pull request rewrites the spec to follow the current arc, and
repairs the log capture at both ends
> - The benefit is that nightly and beta releases are unblocked, and the
next failure arrives with the logs attached
## Linked Issues or Issue Description
No existing issue. Describing it inline, following
`.github/ISSUE_TEMPLATE/bug_report.yml`.
Refs #12274 (removed the company-naming step from the wizard).
Refs #12135 (the previous alignment of this spec, before #12274).
Refs #12316 (open; also edits `scripts/docker-onboard-smoke.sh`, in the
bootstrap helpers rather than the container lifecycle, so the two
changes do
not overlap. Whichever lands second should rebase and re-run).
**What happened?**
The release smoke fails.
`tests/release-smoke/docker-auth-onboarding.spec.ts`
never gets past its first wait:
```
✘ tests/release-smoke/docker-auth-onboarding.spec.ts:43:3 › Docker authenticated onboarding smoke › logs in, completes onboarding, and hires the lead agent
Error: expect(locator).toBeVisible() failed — element(s) not found (timeout 20000ms)
> 33 | await expect(wizardHeading.or(startButton)).toBeVisible({ timeout: 20_000 });
```
The spec waits for an `h3` reading "Name your organization" or a
"Start Onboarding" button. Neither exists. #12274 removed the
company-naming
step; the string now survives only in a code comment and in
`ui/src/components/OnboardingWizard.step.test.tsx`, which asserts it is
*absent*. The steps after the first wait are stale too: the CTA on step
1 is
"Continue" and not "Next", the organization input's placeholder changed,
and
the agent step's `#onboarding-agent-role` picker is gone, so every
onboarding
hire is filed under the neutral `general` role.
The suite runs only from the release pipeline, so nothing on a pull
request
saw the drift. Both `smoke_nightly` and `smoke_beta` call the same
reusable
workflow, so every nightly and every beta was blocked.
The failure also arrived without diagnostics. The job's "Capture Docker
logs"
step is `if: always()`, but it is guarded on `SMOKE_CONTAINER_NAME`,
which the
"Launch Docker smoke harness" step writes to `$GITHUB_ENV` only *after*
the
harness returns. On any failure before that the guard is false, the step
does
nothing, and the upload reports "No files were found". Below that,
`scripts/docker-onboard-smoke.sh` starts the container with
`docker run -d --rm`, so the `docker stop` in its EXIT trap deletes the
container and its logs together — and a container that crashes on its
own is
removed the instant its process exits.
**Expected behavior**
The spec walks the onboarding arc the app actually presents, and proves
the
company is created, the lead agent is hired, and the first task is
seeded and
dispatched. When the smoke fails, the run's artifact carries the
container's
logs.
**Steps to reproduce**
1. Run the Release Smoke workflow against a published artifact that
carries
#12274, or run it locally:
`PAPERCLIPAI_VERSION=2026.828.0-canary.3 SMOKE_DETACH=true
./scripts/docker-onboard-smoke.sh`
2. Run `pnpm run test:release-smoke` against that container.
3. The single spec fails at `openOnboarding()` after 20 seconds.
4. In CI, open the run's `release-smoke` artifact. It has no
`docker-onboard-smoke.log`.
**Paperclip version or commit**
`2026.828.0-canary.3` (commit 8316ceb0b).
**Deployment mode**
Docker.
**Installation method**
npm / pnpm global install (the container runs `npx
paperclipai@<version>`).
**Node.js version**
v24.20.0 inside the container.
**Relevant logs or output**
```
Running 1 test using 1 worker
✓ 1 [chromium] › tests/release-smoke/docker-auth-onboarding.spec.ts:76:3 › Docker authenticated onboarding smoke › logs in, completes onboarding, and hires the lead agent (7.0s)
1 passed (8.7s)
```
That is the result after this change. Before it, the same command failed
at
the first wait, as quoted above.
## What Changed
- `tests/release-smoke/docker-auth-onboarding.spec.ts` now follows the
current
arc. It signs in, opens `/onboarding`, names the organization and
presses
"Continue" (which creates the company and routes straight to the agent
step,
because onboarding no longer asks for a mission), names the lead and
presses
"Next", presses "Connect" on the default adapter to hire, then presses
"Get started" to launch.
- The spec addresses controls by role and accessible name, or by id
where one
exists (`#onboarding-agent-name`). Step 1's field has no id and no
associated
label, so it is found as the wizard's only text box rather than by its
placeholder copy.
- The spec asserts the hired agent's role is `general`, which is what
the arc
files every onboarding hire under. Every other API assertion is
unchanged.
- The spec navigates to `/onboarding` explicitly and drops any saved
onboarding
draft first, so it can run twice against one instance. The suite retries
once
in CI. It still asserts that a company-less board routes sign-in into
onboarding, guarded on the board actually being empty.
- `scripts/docker-onboard-smoke.sh` accepts `SMOKE_CONTAINER_NAME`,
drops
`--rm`, removes the container itself, and dumps `docker logs` to
`SMOKE_LOG_FILE` before the teardown.
- `.github/workflows/release-smoke.yml` pins the container name in the
job's
`env`, so every `always()` step has it before anything runs. The capture
step
refreshes the log from a live container when there is one, keeps the
harness's dump when there is not, and writes a one-line explanation when
there is neither. The upload's paths are literals, and
`if-no-files-found: error` makes a broken diagnostics path fail rather
than
warn.
- `scripts/docker-onboard-smoke.test.mjs` pins that wiring. It is added
to
`test:release-registry`, which runs on every pull request.
- `doc/DOCKER.md` documents `SMOKE_CONTAINER_NAME` and `SMOKE_LOG_FILE`.
## Verification
The spec was run against a real container built from the published
`2026.828.0-canary.3` artifact, exactly as the workflow runs it.
```sh
SMOKE_CONTAINER_NAME=release-smoke-onboard \
HOST_PORT=3232 DATA_DIR=<tmp>/smoke-data \
PAPERCLIPAI_VERSION=2026.828.0-canary.3 \
SMOKE_READY_TIMEOUT_SECONDS=420 SMOKE_DETACH=true \
SMOKE_METADATA_FILE=<tmp>/release-smoke.env \
SMOKE_LOG_FILE=<tmp>/docker-onboard-smoke.log \
./scripts/docker-onboard-smoke.sh
PAPERCLIP_RELEASE_SMOKE_BASE_URL=http://localhost:3232 \
PAPERCLIP_RELEASE_SMOKE_EMAIL=smoke-admin@paperclip.local \
PAPERCLIP_RELEASE_SMOKE_PASSWORD=paperclip-smoke-password \
PAPERCLIP_PLAYWRIGHT_CHANNEL=chrome \
pnpm run test:release-smoke
```
```
Running 1 test using 1 worker
✓ 1 [chromium] › tests/release-smoke/docker-auth-onboarding.spec.ts:76:3 › Docker authenticated onboarding smoke › logs in, completes onboarding, and hires the lead agent (7.0s)
1 passed (8.7s)
```
The same command was run a second time against the same, now non-empty,
instance. That covers the retry path, and it also passes.
The log capture was verified by making the container die during startup:
```sh
PAPERCLIPAI_VERSION=0.0.0-no-such-version \
SMOKE_CONTAINER_NAME=release-smoke-onboard SMOKE_LOG_FILE=<tmp>/fail.log \
./scripts/docker-onboard-smoke.sh
```
`<tmp>/fail.log` was written and carried the cause:
```
npm error code ETARGET
npm error notarget No matching version found for paperclipai@0.0.0-no-such-version.
```
The container was removed afterwards. On `master` this file is never
written,
because `--rm` deletes the container the moment its process exits.
The workflow's capture step was run by hand against three states: a live
container (258 lines), a removed container with the harness's dump
already on
disk (258 lines kept), and neither (a one-line explanation).
Unit coverage:
```sh
pnpm run test:release-registry # 93 tests, 93 pass
```
Nothing under `ui/` changed, so `pnpm --filter @paperclipai/ui
typecheck` was
not required. `tests/release-smoke` is outside the TypeScript project
references; Playwright compiles it at run time, which the runs above did
three
times.
## Risks
Low risk. Nothing ships to users. The change touches one Playwright
spec, one
smoke script, and one workflow.
Points worth a reviewer's attention:
- **This suite gates every nightly and beta, and it runs only
post-merge.**
`smoke_nightly` and `smoke_beta` both call `release-smoke.yml`, and no
pull
request runs it. Drift between the wizard and this spec is therefore
invisible until a release is already blocked, which is how this bug
reached
a release train. I think the arc deserves an earlier check. The cheapest
version is the one added here: `scripts/docker-onboard-smoke.test.mjs`
runs
on every pull request and pins the harness wiring. The full container
smoke
is too slow for the pull request path, but a UI-level test of the arc's
step
sequence would catch exactly this class of drift, and
`ui/src/components/OnboardingWizard.step.test.tsx` is already the right
home for it. I did not add it here, to keep this change to the repair.
- **Dropping `--rm`.** The container is now removed by the script's
cleanup
instead of by Docker. The script already ran `docker rm -f` before
starting,
and the workflow's final step removes it too, so a leaked container is
cleaned up on the next run either way. A developer who kills the script
with
`SIGKILL` will leave a stopped container behind, where previously they
would
not.
- **`if-no-files-found: error` on the upload.** The capture step now
always
writes the log file, so the upload always has at least one path to
match. If
that ever stops being true, the job fails instead of warning. That is
deliberate.
- **The spec drops the saved onboarding draft before it walks.** A stale
draft
makes step 1 skip company creation and hire into the previous run's
company.
That state only exists when the spec runs twice against one instance. A
fresh
release-smoke container never has it.
## Model Used
Claude (Anthropic), Claude Opus, 1M context, extended thinking, agentic
tool
use via Claude Code. The container, the Playwright runs, and the failure
injection were driven as real commands on a local Docker host.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app that people use to manage AI agents
for work.
> - Paperclip applies database migrations in numeric order.
> - Two branches can create the same migration number before either
branch merges.
> - The existing repository check can find duplicates only after both
histories are present in one checkout.
> - A pull request must compare its new migrations with the target
branch before merge.
> - This pull request adds that comparison to the existing PR policy
job.
> - The benefit is an early failure with exact renumbering instructions.
## Linked Issues or Issue Description
**What existing behavior does this improve?**
The PR policy check for files in `packages/db/src/migrations`.
**Current behavior**
A stale branch can add the same migration number as the target branch.
The existing check does not compare PR additions with the target branch
migration tip.
**Proposed behavior**
The policy job fails when a new PR migration number is not greater than
every migration on the target branch. The error names the conflict, the
next safe number, and the related files to update.
**Reason and benefit**
This prevents duplicate or out-of-order migration numbers from reaching
`master`. It also gives contributors and agents a direct repair
procedure.
**Breaking changes**
None. The change rejects migration numbering that is already unsafe.
## What Changed
- Added a dependency-free check that compares new PR migration files
with the target branch tip.
- Added the check to the existing PR policy job.
- Added tests for no-op, valid, duplicate, and lower-number cases.
## Verification
- `node --test '.github/scripts/tests/*.test.mjs'` passed 133 tests.
- `pnpm -r typecheck` passed.
- `pnpm build` passed.
- `pnpm test:run` passed 4,889 tests and failed 24 unrelated macOS path
and wildcard-listener tests that also affect the current `master`
checkout.
## Risks
- Low risk. The check reads Git history and does not modify migrations.
- The check permits gaps. It only requires each new migration number to
follow the target branch tip.
- The existing migration check continues to validate duplicate numbers,
snapshots, and journal entries inside the PR.
## Model Used
OpenAI Codex, GPT-5. The exact serving model ID and context-window size
are not exposed in this session. Reasoning, tool use, web access, and
code execution were enabled.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Paperclip supports opt-in Sentry error monitoring for server and
browser errors.
> - The hosted image must include the server package when an operator
sets SENTRY_DSN.
> - The server package is an optional peer in the source tree, so the
image did not include it.
> - This pull request installs the declared server package in the hosted
image and checks the result.
> - The benefit is a hosted tenant can send server errors without a
manual package install.
## Linked Issues or Issue Description
No public issue exists for this change.
**What happened?**
The hosted image did not include the declared @sentry/node server
package. A hosted tenant could set SENTRY_DSN, but the server could not
load the package from the image.
**Expected behavior**
The hosted image must include the exact @sentry/node version from
server/package.json. The self-hosted image must remain without this
optional package.
**Steps to reproduce**
1. Build or pull the hosted image.
2. Resolve @sentry/node from the server package path.
3. Compare its version with server/package.json.
4. Confirm that the tsx loader path still resolves.
**Paperclip version or commit**
Commit b6ff556a33ebdbe764b7f495951cd59009776608.
**Deployment mode**
Docker hosted image.
## What Changed
- Add a cloud-server-deps Docker stage that installs the declared
@sentry/node version in isolation.
- Copy the isolated package into the cloud image without changing the
production image.
- Add a probe that checks the tsx loader and the resolved Sentry
version.
- Run the probe after the hosted image push in the Docker workflow.
- Add server tests and update the observability documentation.
## Verification
- Run `pnpm exec vitest run --project @paperclipai/server
server/src/__tests__/cloud-image-sentry.test.ts`.
- Confirm that the changed test passes in CI.
- Confirm that all pull request checks pass.
- Note that the Docker workflow does not run for pull requests. It runs
after a push to master, for configured tags, or after manual dispatch.
## Risks
- Low risk. The production image body stays unchanged.
- The cloud image adds the declared Sentry package and a small
dependency tree.
- The workflow probe fails if the image loses the tsx loader or resolves
a different Sentry version.
## Model Used
OpenAI GPT-5; exact model version supplied by the execution service;
tool use and code execution; context window not specified.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` /
`Closes: #` / `Refs: #` OR (b) described the issue in-PR following the
relevant issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Agent runs execute inside the server container, and they spawn many
short-lived descendants: git, the adapter CLI, esbuild, sh
> - The server image sets `ENTRYPOINT ["docker-entrypoint.sh"]`, and
that entrypoint ends in `exec`, so node becomes PID 1
> - Node reaps only the children it spawned itself. It installs no
`SIGCHLD`/`waitpid` handler for orphans that the kernel re-parents onto
PID 1, so those orphans stay as zombies forever
> - Zombies accumulate monotonically. When the cgroup pid limit is
reached, every `fork()` in the container fails and the instance is dead
> - This pull request installs `tini` and makes it PID 1 in front of the
existing entrypoint, adds a behavioural test that proves reaping, and
adds a `pids_limit` backstop to both compose files
> - The benefit is that a long-running container no longer degrades into
total fork failure, and a future regression is caught by CI instead of
by an outage
Depends-on: none — this change is self-contained in the image build and
its tests, and it touches no other in-flight branch
## Linked Issues or Issue Description
No public GitHub issue exists for this defect. It was found on a live
long-running instance. Description follows the bug report template.
**What happened?**
The server container ran for 22 hours and reached 2039 of 2048 pids in
its cgroup. Of 1760 processes, 1731 were zombies, and all 1731 had PID 1
as their parent. PID 1 was `node --import
./server/node_modules/tsx/dist/loader.mjs server/dist/index.js`. Zombies
accrued at about 79 per hour and were never reaped. The oldest zombie
was 20.8 hours old against a container uptime of 22.0 hours, so nothing
had been reaped since boot. Once the pid limit was reached, `git` and
`gh` failed with `pthread_create failed: Resource temporarily
unavailable`.
**Expected behavior**
PID 1 reaps orphaned processes that the kernel re-parents onto it. The
pid count of a long-running container stays flat instead of growing
without bound.
**Steps to reproduce**
1. Start the server image without `docker run --init` and without `init:
true`.
2. Run agent work that spawns descendants which outlive their immediate
parent.
3. Read `/sys/fs/cgroup/pids.current` and count processes in `Z` state
over several hours.
4. The zombie count grows monotonically and every zombie has PPID 1.
**Relevant logs or output**
```
cgroup pids.current / pids.max : 2039 / 2048
total processes : 1760
zombies : 1731 (98.4%)
parent of every zombie : PID 1 (1731/1731)
PID 1 cmdline : node --import .../tsx/dist/loader.mjs server/dist/index.js
container uptime : 22.0 h
oldest zombie : 20.8 h median: 14.4 h
zombie names : git 717, claude 280, MainThread 167, sleep 141,
esbuild 138, postgres 76, sh 65, sccache 50
```
**Additional context**
The fix pattern is already in this repository.
`docker/agent-runtime/Dockerfile.base` installs `tini` and sets
`ENTRYPOINT ["/usr/bin/tini", "--"]`. It was never applied to the server
image.
## What Changed
- `Dockerfile`: install `tini` in the `base` stage and set `ENTRYPOINT
["/usr/bin/tini", "--", "docker-entrypoint.sh"]`. The entrypoint stays
in the exec chain, so UID/GID remapping, `gosu`, and graceful shutdown
are unchanged.
- `scripts/assert-orphan-reaping.sh` (new): a behavioural probe. It
spawns a leader that forks a grandchild, exits the leader, and asserts
that the orphaned grandchild leaves `Z` state instead of persisting. It
fails closed if the grandchild is not re-parented onto PID 1, so a pass
cannot mean the check ran too early.
- `.github/workflows/docker.yml`: run that probe against the pushed
image after the publish step. The publish step is multi-arch with `push:
true`, so nothing is loaded into the runner daemon and the pushed tag is
the only thing to test. The cloud variant is `FROM production` and
inherits the same `ENTRYPOINT`.
- `scripts/docker-build-test.sh`: run the same probe against a local
build.
- `docker/docker-compose.yml` and
`docker/docker-compose.quickstart.yml`: add `pids_limit: 2048` as a
backstop, so a future leak dies visibly at its own ceiling instead of
starving the host of pids.
- `server/src/__tests__/container-init-reaping.test.ts` (new): 13
assertions that guard the configuration the probe depends on.
No per-orchestrator init lever was added. The image owning PID 1 covers
compose, plain `docker run`, the quadlet units, and the ECS task
definition in one place. Adding `init: true` in compose or
`initProcessEnabled` on the ECS task would nest a second init around
`tini`, and `tini` then warns on every boot that it is not PID 1. The
new test asserts the absence of both levers across all three manifests,
so the decision survives the next edit.
## Verification
| Check | Result |
|---|---|
| `scripts/assert-orphan-reaping.sh` against a real init | Grandchild
re-parented to PPID 1, then reaped. Exit 0. |
| Same probe forced against a genuine zombie | Reports `Z` and fails.
The failure branch is not vacuous. |
| Config guard against the pre-fix files | Exactly the 3 relevant
assertions turn red. |
| Config guard with `tini` removed from `apt-get` but the comments kept
| Red. It checks the install, not a mention of the name. |
| `cd server && npx vitest run
src/__tests__/container-init-reaping.test.ts` | 13 passed |
| `npx tsc --noEmit -p server` | Clean |
| `node scripts/check-docker-deps-stage.mjs` | PASS |
| `node --test scripts/release-verify-workflow.test.mjs` | 8 passed |
Not verified locally: no container runtime is available in the authoring
environment, so the probe has not run against a build of this image. The
new `docker.yml` step runs it against the pushed image on this PR.
## Risks
Low risk, but it is an image and entrypoint change, so it affects
deployments.
- `tini` adds one small package to the `base` stage.
`docker/agent-runtime/Dockerfile.base` already installs it from the same
Debian archive.
- Signal handling changes shape: `tini` receives `SIGTERM` and forwards
it to the entrypoint, which `exec`s node. `tini` forwards signals to its
direct child by default, and the exec chain keeps node as that child, so
graceful shutdown is preserved. A reviewer should confirm this on a real
stop.
- `pids_limit: 2048` is new for compose users. A deployment that
legitimately needs more than 2048 processes would now hit the ceiling.
The measured steady state on a busy instance was under 400.
- If a deployment already passes `--init` or `init: true`, `tini` runs
under another init and prints a warning that it is not PID 1. Reaping
still works because the outer init handles it. The compose files in this
repository do not set `init: true`.
## Model Used
Claude Opus 5 (`claude-opus-5`), extended thinking, with tool use and
code execution in an agent harness.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local issues or links
- [x] My branch name describes the change and contains no internal
ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: zannis <1011451+zannis@users.noreply.github.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The runner package now has protocol, transport, provider, catalog,
and authorization foundations.
> - Its first upstream package boundary should expose only the
implemented runtime and test-helper surfaces.
> - Rust correctness belongs in the repository existing build
verification, without introducing a parallel release process.
> - Direct package creation must build the files declared by the package
manifest.
> - This pull request defines the minimal package API and verifies the
optimized runner binaries in the existing PR and release Build jobs.
> - The benefit is a production-ready runner package boundary with
minimal build-process change.
## Linked Issues or Issue Description
Refs #11962
This pull request replaces one bounded part of the archived large runner
change. It follows the package-local authorization change in #12126.
## What Changed
- Export only `@paperclipai/paperclip-runner` and
`@paperclipai/paperclip-runner/testing`.
- Keep Node-only fixture loading and semantic conformance helpers out of
the runtime root.
- Add a provider-neutral semantic conformance kit with stable JSON
comparison and fail-closed input checks.
- Keep deferred SDK, eval, browser, React, lab, and command surfaces
private.
- Pin the runner Rust toolchain to 1.97.1 with the minimal profile and
`rustfmt`.
- Run the Rust workspace tests in release mode.
- Launch the optimized `paperclip-runnerd` and fake-harness binaries in
process-level integration coverage.
- Add one `pnpm --filter @paperclipai/paperclip-runner check:all` step
to each existing PR and release Build job.
- Make the existing server `prepack` lifecycle run its existing build
after it prepares UI assets.
- Document that no production adapter starts runnerd yet.
This revision adds no standalone GitHub Actions job. It adds no server
runner dependency or runner vendoring. It adds no Docker bootstrap or
clean-consumer harness. It does not change `pnpm-lock.yaml`.
## Verification
- `pnpm --filter @paperclipai/paperclip-runner check:all`
- 66 TypeScript tests
- 8 protocol contract tests
- 56 Rust unit and integration tests
- Release-mode integration coverage launches the optimized runnerd and
fake-harness binaries.
- `pnpm --filter @paperclipai/server exec vitest run
src/__tests__/server-package-build-script.test.ts` (2 tests)
- Clean `pnpm pack` from `server/` rebuilt the server and produced both
`package/dist/index.js` and `package/dist/index.d.ts`.
- `node --test scripts/__tests__/release-verify-workflow.test.mjs` (8
tests)
- `pnpm -r typecheck`
- `pnpm build`
- `pnpm check:token-gates`
- `git diff --check`
- No `pnpm-lock.yaml` diff.
- The diff changes 12 files.
## Risks
The runner adds Rust work to the existing Build jobs. These jobs can
take longer on a cold cache. The pinned toolchain makes contributor and
CI behavior reproducible. Cargo tests use `--release` to verify
optimized executables. The server prepack lifecycle now performs the
build that its published entry points require. This can make direct
server packing slower. This pull request does not wire runnerd into the
server. It does not select runnerd for any adapter. Existing application
execution and finalization paths remain unchanged.
## Model Used
OpenAI Codex with GPT-5. Agentic coding mode used repository tools, code
execution, and automated tests.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release pipeline gates each nightly and beta on a smoke suite
that onboards the published npm artifact and drives the golden path
> - That smoke runs onboarding inside a Docker container, and containers
have no service manager, so the background-service leg of onboarding has
zero automated coverage
> - v2026.824.0 shipped a service install that crash-looped on a missing
shim, and every smoke check stayed green (#12148 fixed the defect
itself)
> - This pull request adds a `smoke_service` job that runs the same
published artifact directly on the runner VM's systemd and requires the
installed service to end up serving
> - The benefit is that a release with a broken service install can no
longer pass the release smoke suite
## Linked Issues or Issue Description
Refs #12148 — the fix for the defect this coverage gap let through. The
gap: the release smoke runs `onboard` with `--yes` inside Docker, which
both skips the service prompt and lacks systemd, so no CI job ever
executed `manager.install()` against a real service manager.
## What Changed
- New `scripts/service-onboard-smoke.sh`: onboards the published
artifact with `--yes --install-service` on a systemd host, then fails
unless the managed shim exists and is executable, `paperclipai.service`
is active, and `/api/health` answers. A health response while the unit
is not active also fails, because that is the signature of something
other than the service serving. The script refuses to run over an
existing managed install unless `SMOKE_FORCE=true`, and cleans up after
itself by default so it is safe to run locally.
- New `smoke_service` job in `.github/workflows/release-smoke.yml`:
starts a user systemd session on the hosted runner (`loginctl
enable-linger` + exported `XDG_RUNTIME_DIR`/`DBUS_SESSION_BUS_ADDRESS`),
runs the script against `inputs.paperclip_version`, and uploads
`systemctl status` + journal output as diagnostics.
- No `release.yml` changes needed: `smoke_nightly` and `smoke_beta` call
this reusable workflow, and a `workflow_call` result aggregates all
jobs, so the new job gates nightly promotion automatically.
## Verification
- `bash -n scripts/service-onboard-smoke.sh` passes and the workflow
YAML parses.
- End-to-end: dispatched this branch's Release Smoke workflow against
the published canary that contains #12148; the `smoke_service` job
onboards, installs the service, and verifies the service serves health.
(Run link in PR comments.)
- Negative case: the same assertions fail against v2026.824.0 —
reproduced in a systemd container during the #12148 investigation: shim
missing, unit in a 203/EXEC restart loop.
## Risks
- Low risk to the product: no application code changes.
- Pipeline risk: a flaky user-session setup on the hosted runner would
block nightly promotion. Mitigated by validating the job end-to-end from
this branch before merge, a 30-minute job timeout, and diagnostics
uploaded on every run.
- The service leg only covers systemd. launchd (macOS) still has no CI
coverage; a macOS runner job is a possible follow-up.
## Model Used
- Claude Fable 5 (Anthropic, model ID `claude-fable-5`), extended
thinking, agentic tool use via Claude Code.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release workflow drafts the upcoming stable's notes skeleton the
moment a beta publishes
> - That skeleton was a bare list of commit subjects, so the notes only
reached the shipped stable's depth after a later authoring pass during
the soak
> - Stable release notes are consistently verbose and thorough; the
initial draft should start that way too
> - This pull request nests each referenced PR's own summary under its
subject line at creation time, and states the density bar in the
authoring skill
> - The benefit is a thorough raw document from day one of the soak,
with no LLM tokens in Actions
## Linked Issues or Issue Description
**What existing behavior does this improve?**
The `draft_stable_notes` skeleton generated at beta publish
(`scripts/draft-stable-notes.sh`).
**Current behavior**
The skeleton groups bare commit subjects by conventional-commit type.
All substance arrives later, when a maintainer or agent rewrites it —
reviewed maintainer feedback: stable notes are a lot more verbose, and
the initial beta notes should be consistent with that.
**Proposed behavior**
Each subject that references a PR carries that PR's own summary nested
beneath it — the PR template's "What Changed" bullets, else the first
prose lines — fetched best-effort via `gh` and skipped silently when
unavailable. The release-changelog skill now states the density bar
explicitly: the beta-keyed draft ships verbatim as the stable's notes
and is written at the previous stable's depth from the first pass.
**Reason and benefit**
The notes author starts from a thorough raw document instead of a commit
list, and beta-time notes match the verbosity the stable will ship with.
## What Changed
- `scripts/draft-stable-notes.sh`: `enrich_pr` nests PR summaries under
subjects; best-effort (`gh` failure or
`DRAFT_NOTES_SKIP_PR_ENRICHMENT=1` degrades to today's output);
pipefail-safe when a "What Changed" section has no bullets.
- `.github/workflows/release.yml`: the `draft_stable_notes` step gets
`GH_TOKEN` so `gh` can read PR bodies.
- `.agents/skills/release-changelog/SKILL.md`: "write at full stable
depth from the first pass" guideline.
- `scripts/draft-stable-notes.test.mjs`: three new tests — enrichment
rendering via a fake `gh`, silent degradation without one, and the
sparse-body case that previously killed the script under `set -o
pipefail`.
## Verification
- `node --test scripts/draft-stable-notes.test.mjs` — 11 pass.
- Live run against the real repository for the current beta
(`2026.818.0-beta.1`, 172 commits): exit 0, 439 nested summary lines;
spot-checked entries carry the correct PRs' What Changed bullets.
- `bash -n` on the script; `release.yml` re-parsed as YAML.
## Risks
- Low: the publish path is untouched; enrichment is read-only `gh` calls
in the post-publish draft job and degrades to the current skeleton on
any failure. Roughly one API call per commit in the range (~170 today) —
well inside the token's rate budget, adds a couple of minutes to a job
with a 10-minute timeout.
## Model Used
Claude Fable 5 (Claude Code)
## Pre-submission checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work, and it takes contributions from outside the core team
> - Those contributions arrive as PRs, and this repository squash-merges
every one of them
> - A squash collapses the whole branch into a single commit authored by
whoever pressed the button
> - So when a maintainer rebases and lands a contributor's stale PR, the
contributor's name survives only if the squash message carries a
`Co-Authored-By` trailer
> - Nothing prompts for that trailer, and the PR page keeps showing the
original author either way, so losing it is invisible at the moment it
happens
> - This pull request has commitperclip detect the situation and print
the exact trailers to paste
> - The benefit is that keeping an outside contributor's name is a
default rather than something a maintainer has to remember
## Linked Issues or Issue Description
No public issue exists. The problem follows, and it is not hypothetical.
**What happened?**
#11370, #11371 and #11379 landed @stubbi's work yesterday. Each of those
PRs carries a comment from me telling them their authorship would be
preserved. All three squash commits went in without a `Co-Authored-By`
trailer, so `git log` credits none of them:
| commit | landed from | credited |
| --- | --- | --- |
| `66515582e` | #9900 | Claude only |
| `bc0b5a164` | #9501 | Claude only |
| `35a9b9873` | #8982 | Claude only |
| `6542ad1f4` | #11259 | ✅ Jannes Stubbemann + Claude |
The last one has the trailer because that message was written by hand
with the contributor in mind. The only difference between the two
outcomes was memory. Master history cannot be rewritten, so those three
are now credited by comment on the original PRs — which is a worse
record than a commit trailer, and the reason to make this automatic.
**Expected behavior**
When a branch carries commits by someone other than the PR author, the
merger is told what trailers the squash needs.
**Paperclip version or commit**
`master` at `92047cac4`.
## What Changed
- `.github/scripts/check-pr-coauthors.mjs` — new gate.
- `.github/scripts/run-quality-gates.mjs` — fetches the PR's commits and
runs it.
- `.github/scripts/tests/check-pr-coauthors.test.mjs` — 12 cases.
- `.github/workflows/pr.yml` — runs `.github/scripts/tests/`.
### Informational, not a failure
The squash message does not exist while the PR is open. This can neither
be verified there nor fixed there, so failing a PR on it would block
work on something its author cannot satisfy. The gate notices that the
situation applies and prints the lines to paste.
Run against #11370's actual commits it produces exactly what was
missing:
```
This branch carries commits by stubbi. Squash-merging drops that authorship
unless the squash message carries their trailers, and nothing else will notice
if it does not. Add to the squash body when merging:
Co-Authored-By: Jannes Stubbemann <stubbi@users.noreply.github.com>
```
### Edge cases it handles
Bots skipped; the PR author's own commits skipped; logins compared
case-insensitively (`PR_AUTHOR` does not always arrive in the same case
as the commit author login); each contributor listed once however many
commits they wrote; and a commit GitHub could not match to an account
falls back to its raw git author — that identity being the one most
likely to be lost, not least likely.
Paging stops at the API's own 250-commit ceiling rather than spinning on
full pages of nothing new.
### The test directory was not running
`.github/scripts/tests/` held ten test files covering the existing
gates, and no workflow ran any of them. Adding an eleventh would have
meant adding a test that never executes, so `pr.yml` now runs the
directory. All **149** pass, including the 137 that were already there
and previously unverified in CI.
## Verification
- 149 tests pass via `node --test '.github/scripts/tests/*.test.mjs'` —
the exact command CI now runs.
- The gate was run against the real commit shape from #11370 and
produces the missing trailer verbatim.
This PR is its own negative control: the branch carries only my commits,
so the new gate should stay silent on it. If commitperclip prints a
co-author note below, the gate is wrong.
## Risks
Low. Informational output only — it cannot fail a PR, and `allPassed` is
unchanged.
It adds one API call per gate run (`/pulls/{n}/commits`), fetched in the
same `Promise.all` as the existing PR and files calls.
Enabling the previously-unrun test directory could in principle surface
a pre-existing failure; all 149 pass locally, so it does not.
Revert the commit to restore.
## Model Used
Claude Opus 5 (`claude-opus-5`), through Claude Code. Extended thinking
enabled. Tool use enabled: file read and edit, shell for test runs.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links
- [x] My branch name describes the change and contains no internal
ticket id
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Paperclip runs across the CLI, server, adapters, plugins, CI, and
container images.
> - These surfaces declared different Node.js versions from 20 through
24.
> - A newer `@types/node` major can expose APIs that the supported
runtime does not provide.
> - Node.js 20 is no longer a suitable project baseline, and Node.js 24
is the current LTS line.
> - This pull request sets Node.js 24.11.0 as one repository-wide
baseline, adds a drift check, and gives users actionable startup
guidance when their runtime is too old.
> - The benefit is one clear runtime contract for development, release,
installation, and published packages.
## Linked Issues or Issue Description
Refs #2734
Refs #11727
Refs #739
## What Changed
- Require Node.js 24.11.0 or newer in all 42 package manifests and
runtime checks.
- Use Node.js 24 in GitHub Actions, Docker images, smoke images, sandbox
setup, portable installs, and esbuild targets.
- Align every direct `@types/node` declaration on `^24.0.0`.
- Prevent Dependabot from opening major `@types/node` upgrades without a
matching runtime decision.
- Add `.nvmrc` and a CI policy check for Node version drift.
- Update ACP version gates, tests, and user documentation for the new
minimum.
- Print a non-blocking warning on CLI and server startup when Node is
unsupported, with remediation through a version manager or the
documented downloaded `install.sh` workflow.
- Deduplicate that warning when `paperclipai run` boots the CLI and
server in the same process.
## Verification
- `node scripts/check-node-version-policy.mjs`
- `node --check scripts/check-node-version-policy.mjs`
- `node --check cli/esbuild.config.mjs`
- `node --check scripts/generate-npm-package-json.mjs`
- `bash -n scripts/install.sh scripts/test-install-sh-docker.sh
scripts/e2e-install-lifecycle.sh`
- Parsed all 42 package manifests and confirmed `engines.node` is
`>=24.11.0`.
- `git diff --check`
- `vitest run
packages/adapter-utils/src/sandbox-install-command.test.ts` passed with
3 tests.
- `vitest run cli/src/node-version.test.ts` passed with 4 tests.
- Directly exercised the shared warning helper for unsupported-version
messaging and same-process deduplication.
- The focused exe.dev suite could not resolve the locally unbuilt plugin
SDK from this isolated worktree. A full offline workspace install was
also blocked because the package-manager signature verifier requires
registry access. The full suite was not run locally; draft CI performs a
clean install and evaluates the wider impact.
## Risks
- This is a breaking runtime change for users, plugins, and deployments
that still use Node.js 20 or 22.
- Published workspace packages will now produce an engine warning or
failure in strict package managers on older Node.js releases.
- Node.js 24 can reveal dependency, native module, Playwright, or agent
CLI compatibility issues in CI.
- The bootstrap installer now installs Node.js 24 when the current
runtime is older than 24.11.0.
- The portable sandbox fallback is pinned to Node.js 24.11.0 and depends
on that upstream tarball remaining available.
- Unsupported runtimes continue booting after a warning, so a later
incompatibility can still fail at its point of use.
- The CLI and server share the warning policy through the published
`@paperclipai/shared` package; packaging checks must keep that subpath
export available.
- This PR does not commit `pnpm-lock.yaml` because repository policy
assigns lockfile generation to CI.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
- OpenAI Codex based on GPT-5. The exact deployment ID and context
window are not exposed in this session. Reasoning, repository tools,
shell execution, and GitHub tools were enabled.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release workflow publishes every public package to npm and polls
each version's registry visibility before continuing
> - #11834 raised the poll budget from 60 seconds to 5 minutes after npm
CDN propagation lag failed four canary runs
> - The very next canary run missed by ten seconds:
`adapter-opencode-local@2026.821.0-canary.2` was accepted at 07:02:27
UTC and became visible at 07:07:40, just past the final poll
> - This pull request doubles the per-package budget to 10 minutes
> - The benefit is a release train that survives the one consistently
slow package; the 90-minute publish job timeout from #11834 already
absorbs it
## Linked Issues or Issue Description
Follow-up to #11834. Evidence in the `Release` run for `16149a75f`:
every package's publish became visible within seconds except
`adapter-opencode-local`, which has lagged 3-5+ minutes on all of
today's runs and exceeded the 5-minute budget by ten seconds on the
latest.
## What Changed
- `NPM_PUBLISH_VERIFY_ATTEMPTS` 30 → 60 (with
`NPM_PUBLISH_VERIFY_DELAY_SECONDS: "10"`, a 10-minute per-package
budget), plus the comment documenting the observed near-miss.
## Verification
- Same env-override plumbing verified in #11834; only the numeric budget
changes. The next master push (this merge) exercises the canary path.
## Risks
- Low risk: a genuinely failed publish reports in up to 10 minutes;
healthy publishes exit the poll on first visibility.
## Model Used
Claude Fable 5 (Anthropic, `claude-fable-5`) with extended thinking and
agentic tool use via the Claude Code CLI.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable (not applicable:
numeric workflow env tuning)
- [x] I have updated relevant documentation to reflect my changes
(workflow comment updated)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release workflow publishes every public package to npm per
master push (canary) and per promotion channel
> - `release.sh` polls the registry after each publish and aborts the
whole release when a version is not visible within 60 seconds
> - npm accepts publishes immediately, but its CDN can lag packument
propagation by several minutes; on 2026-08-21 this failed four
consecutive canary runs mid-loop even though every publish succeeded
> - This pull request sets the script's existing visibility-budget env
overrides at the workflow level to 10 minutes
> - The benefit is a release train that tolerates registry propagation
lag; a healthy publish still exits the poll on its first visible check
## Linked Issues or Issue Description
Not applicable for a `ci:` workflow tuning change. Evidence: four
consecutive `Release` runs on master failed in `publish_canary` with
"npm did not publish and expose <package>@<version>", while the raw
registry packument shows each of those versions present minutes later
(`2026.821.0-canary.0` accepted 01:36 UTC, visible 01:40;
`2026.821.0-canary.1` accepted 05:54, visible 05:57).
## What Changed
- Set `NPM_PUBLISH_VERIFY_ATTEMPTS: "30"` and
`NPM_PUBLISH_VERIFY_DELAY_SECONDS: "10"` in the `Release` workflow's
top-level `env`, raising `release.sh`'s post-publish visibility poll
from 60 seconds to 5 minutes per package for every channel. Both
variables are existing overrides read by the script
(`scripts/release.sh` lines 311-312); no script change.
- Raised the four publish jobs' `timeout-minutes` from 45 to 90 so
several laggard packages fit inside the job without exhausting it before
the tag push / Docker / release steps.
## Verification
- `release.sh` reads the two env overrides with defaults
(`${NPM_PUBLISH_VERIFY_ATTEMPTS:-12}` /
`${NPM_PUBLISH_VERIFY_DELAY_SECONDS:-5}`), so workflow-level env reaches
`publish_package_to_npm_and_wait` unchanged.
- Not run: a live release (needs the npm-canary environment). The next
master push exercises the canary path with the new budget.
## Risks
- Low risk: a genuinely failed publish now takes up to 5 minutes to
report instead of 1, and a pathological batch where most packages lag
the full budget still fails inside the 90-minute job — that pattern
means a real registry incident. The poll exits early on success, so
healthy releases are unaffected.
## Model Used
Claude Fable 5 (Anthropic, `claude-fable-5`) with extended thinking and
agentic tool use via the Claude Code CLI (release log forensics against
raw registry packument timestamps).
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable (not applicable:
workflow env tuning)
- [x] I have updated relevant documentation to reflect my changes
(comment in the workflow documents the budget rationale)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip's `commitperclip-review` workflow runs
`check-pr-security.mjs` on every PR and files a **draft security
advisory** whenever one of its heuristics fires (#6469).
> - The heuristics fire on most ordinary PRs: any change under
`server/src/routes/agents.ts` / `companies.ts` / `approvals.ts` /
`authz.ts` / `MarkdownBody.tsx`, any test file containing `fetch(` /
`exec(` / `process.env.X`, any `key: "<20+ chars>"` string (it flagged
`pluginKey: "paperclipai.plugin-llm-wiki"`), any touch of
`.github/workflows/`.
> - The repository now holds **1,566 commitperclip-authored draft
advisories** against ~99 human-reported ones, burying the reports that
matter under the 🔒 Security tab.
> - Nothing consumes them: no code reads the drafts, nothing reads the
`security-review` check run the script also posts, and `master` has no
required status checks. The "Review and dismiss if not a real concern"
footer assumed a human triage loop that never existed.
> - A second bug made it worse: the advisories endpoint is
cursor-paginated and ignores `page=`, so `findExistingDraftAdvisory`
only ever saw the newest 100 drafts and re-flagged PRs got a second
draft (1,566 drafts for 1,386 distinct PRs; 136 PRs have 2+).
> - Removing the gate stops the flood at the source; the quality gates
and Dependency Review carry on unchanged.
## Linked Issues or Issue Description
**Problem:** `check-pr-security.mjs` files a draft security advisory for
nearly every PR, flooding the repository's advisory list with
bot-authored noise that no one reads. Human-reported advisories in
`triage` state are buried among ~1,560 `🚨 Security flag — PR #NNNN`
drafts.
**Expected:** the advisory list contains only real vulnerability
reports. Heuristic PR checks, if wanted at all, do not create disclosure
records.
## What Changed
- Deleted `.github/scripts/check-pr-security.mjs` and
`.github/scripts/tests/check-pr-security.test.mjs`.
- Removed the `Run security gates` step from
`.github/workflows/commitperclip-review.yml`, and the `security-events:
write` permission that only it used.
- No other script imports from the removed module (`resolveBaseRef`
lives in `check-pr-dependencies.mjs` and stays).
## Verification
- `node --test .github/scripts/tests/*.test.mjs` → 114 pass, 0 fail.
- `grep -rn check-pr-security .github` → no remaining references.
- The 1,563 existing bot drafts are being closed out-of-band via the API
(there is no delete endpoint for advisories).
## Risks
- Low. The only behaviour removed is the draft-advisory filing and the
informational `security-review` check run, neither of which is consumed
by code or branch protection.
- Recommended follow-up for an org admin: drop `security_advisories:
write` from the commitperclip App's permissions so no workflow can
recreate this.
## Model Used
Claude Fable 5 (claude-fable-5) via Claude Code, with tool use: GitHub
API reads, file edits, local test runs.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have run tests locally and they pass
- [x] I have considered and documented any risks above
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_016HitAcRu3NW5YDeBxXxePi
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release channel system promotes builds canary → nightly → beta →
stable, and stable releases publish a GitHub Release from
`releases/vYYYY.MDD.P.md`
> - The stable lane requires that notes file to exist inside the
promoted source commit, but the file is named for the promotion date,
which is unknown when the source commit is created
> - A promoted beta can therefore never pass the notes check: every
happy-path stable is forced through the candidate-branch fix path, with
a soak-gate justification, for a notes-only change
> - This pull request drafts the notes automatically when the beta is
published and lets the stable promotion read them from `master`
> - The benefit is a walkable stable happy path: the soak gate stays
exact, notes get a real review window during the soak, and the
justification path returns to its real purpose (cherry-picked fixes)
## Linked Issues or Issue Description
**What existing behavior does this improve?**
The stable promotion path in the release channel system (`release.yml`,
`scripts/release.sh`).
**Current behavior**
`release.sh stable` requires `releases/vYYYY.MDD.P.md` in the
checked-out source tree, and `publish_stable` checks out the exact
promoted SHA. The soak gate requires a `beta/v*` tag to point at that
same SHA. No commit can satisfy both for a promoted beta, so a stable
promotion must cut a candidate branch with a notes-only commit and
bypass the soak gate with a written justification. Release notes are
also written at promotion time, under time pressure, with no review
window.
**Proposed behavior**
When a beta publishes, a `draft_stable_notes` job generates a grouped
notes skeleton at `releases/beta/v<beta-version>.md` and pushes it to a
machine-owned branch; a human opens the PR and edits it during the 3-day
soak. The stable preflight resolves notes before the `npm-stable`
approval gate: source-tree notes first (the candidate fix path,
unchanged), then the merged beta-keyed file on `master`; it fails early
with the missing path named when neither exists. After the stable ships,
a canonicalization job pushes a branch that moves the file to
`releases/vYYYY.MDD.P.md`.
Related (not duplicates): #11006 and #11008 introduced the nightly and
beta lanes this builds on; older changelog PRs (for example #10669)
authored notes manually at promotion time, which is the flow this
replaces.
**Reason and benefit**
The happy path becomes: promote the exact soaked SHA, no justification,
notes reviewed during the soak instead of written at the gate. The
`releases/vYYYY.MDD.P.md` invariant still holds durably via the
canonicalization PR.
## What Changed
- `scripts/release.sh`: new `--notes-file PATH` (stable only) overrides
where the pre-publish notes check looks, so notes can live outside the
source checkout without dirtying the worktree.
- `scripts/create-github-release.sh`: same `--notes-file` override for
the GitHub Release body.
- `scripts/draft-stable-notes.sh` (new): deterministic skeleton
generator — commit subjects from the newest stable tag (falling back to
the previous beta, then full history) to the beta's source commit,
grouped into Features / Fixes / Other.
- `.github/workflows/release.yml`:
- `draft_stable_notes` job after `publish_beta`: runs the generator and
force-pushes `release-notes/v<beta-version>`; the job summary links the
compare page. It recreates the beta tag locally if the tag push was
rejected (the known workflows-permission case), so drafting is not
blocked on manual tag recovery.
- `preflight_stable`: computes the target stable version (`release.sh
stable --print-version`) and resolves the notes source (`source_tree` →
`master_beta` → fail early / warn on dry run); new outputs.
- `publish_stable`: materializes `master`-side notes into `RUNNER_TEMP`
and passes `--notes-file` to both scripts; outputs the published stable
version.
- `canonicalize_stable_notes` job: pushes the `git mv` branch after a
stable that used `master`-side notes.
- `doc/RELEASING.md`, `doc/RELEASE-CHECKLIST.md`: document the
drafted-notes flow, the preflight resolution order, and the
canonicalization step; the LLM changelog flow now targets the draft
branch during the soak.
- `.agents/skills/release-changelog/SKILL.md`,
`.agents/skills/release-changelog-discord-message/SKILL.md`: the
notes-authoring skills now describe this flow — range ends at the beta
source commit (not `HEAD`), the file is beta-keyed on the
`release-notes/v<beta-version>` branch (seeded with
`scripts/draft-stable-notes.sh` for betas that predate the automation),
and the canonicalization link caveat is called out for announcements.
## Verification
- `node --test scripts/draft-stable-notes.test.mjs` — 6 tests, temp
git-repo fixtures: grouping, stable-tag range, previous-beta and
full-history fallbacks, default output path, malformed version, missing
tag.
- `node --test scripts/release-lib.test.mjs` — unchanged suite still
green.
- `bash -n` on both changed shell scripts; `release.yml` re-parsed as
YAML.
- `./scripts/release.sh stable --print-version` unchanged (prints the
next stable version); `--notes-file` on a non-stable channel fails with
a clear error.
- Not exercised end-to-end: the new workflow jobs need a real beta
publish to run. The first beta after merge is the live test; the draft
job is additive and cannot affect the publish result (it runs after
`publish_beta` completes).
## Risks
- Low risk to publishing itself: `--notes-file` defaults preserve
today's behavior everywhere; the draft and canonicalization jobs are
additive and run after the publishes succeed.
- The preflight now fails a real stable run when no notes are found.
That is the intended fail-early behavior (it previously failed later,
inside `publish_stable`, after the `npm-stable` approval).
- `draft_stable_notes` force-pushes only the machine-owned
`release-notes/v<beta-version>` branch; a beta re-cut regenerates it
cleanly.
- The stable version computed at preflight could differ from the
published one if a run crosses UTC midnight between the two jobs; the
materialized notes are passed by path, so the publish still succeeds,
and the canonicalization job uses the actually-published version.
## Model Used
Claude Fable 5 (Claude Code)
## Pre-submission checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
<!-- Write all pull request text in Simplified Technical English
(ASD-STE100). -->
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The PR verify workflow gates every pull request; its wall-clock time
sets the feedback loop for all contributors
> - In a recent successful PR run (actions run 32012408876), the slowest
check was "Verify serialized server suites (1/5)" at 337s, while its
four sibling shards finished in 212-238s
> - The serialized lane assigns suites to shards round-robin over an
alphabetical list, so the heavy heartbeat and issues suites cluster on
one runner
> - The general-server lane already solves this with a duration-aware
LPT partition backed by a recorded manifest
> - This pull request reuses that partitioner for the serialized lane
with a fresh per-suite duration manifest
> - The benefit is a balanced serialized matrix: the measured 968s suite
total levels to about 194s per shard, which removes about 80-100s from
the run's slowest check
## Linked Issues or Issue Description
**What existing behavior does this improve?**
The `Verify serialized server suites` shard matrix in
`.github/workflows/pr.yml` distributes route/authz test suites across
five runners.
**Subsystem affected**
CI / test infrastructure (`scripts/run-vitest-stable.mjs`).
**Current behavior**
`selectSerializedSuites` assigns suites round-robin (`index %
shardCount`) over the alphabetically sorted file list. The heavy suites
cluster on shard 1/5. In actions run 32012408876, shard 1/5 spent 291s
in its test step while the other shards spent 170-201s, which made that
job (337s total) the slowest check of the whole PR run.
**Proposed behavior**
Partition the serialized suites with the same duration-aware LPT
algorithm the general-server lane already uses
(`scripts/general-server-shard.mjs`), backed by a new per-suite duration
manifest. All five shards then carry about 194s of measured test time.
**Reason and benefit**
The slowest check bounds PR feedback time. Balancing the serialized
matrix removes about 80-100s from that bound without adding runners.
**Breaking changes**
None. The partition remains deterministic, complete, and
non-overlapping; suites missing from the manifest get the median weight.
## What Changed
- Added `scripts/serialized-shard-durations.json`: per-suite wall-clock
durations (ms) for all 134 serialized suites, sampled from actions run
32012408876 by diffing consecutive per-suite label timestamps in the
shard logs (captures vitest spawn overhead, not just reported test time)
- `scripts/run-vitest-stable.mjs`: `selectSerializedSuites` now uses the
existing LPT partitioner (`selectGeneralServerShard`) with the new
manifest instead of round-robin
- `scripts/__tests__/run-vitest-stable-shard.test.mjs`: added a
manifest-freshness test and a shard-balance test for the serialized
lane, mirroring the general-server ones
- `.github/workflows/pr.yml`: updated the serialized matrix comment with
the new measurement and mechanism
## Verification
- `node --test ./scripts/__tests__/run-vitest-stable-shard.test.mjs`
passes (13 tests), including the existing test that the serialized
shards form a complete, non-overlapping partition
- Dry-run of all five shards shows estimated totals of
194/194/194/194/193s (round-robin was 276/175/160/172/187s):
`node scripts/run-vitest-stable.mjs --mode serialized --shard-index N
--shard-count 5 --dry-run`
- The `Verify serialized server suites` jobs on this PR run the real
partition end to end
## Risks
- Low risk. Selection logic only; the vitest invocation per suite is
unchanged
- A stale manifest degrades gracefully: unknown suites get the median
weight, and a dedicated test fails if fewer than half the current suites
have recorded durations
## Model Used
- Claude (Anthropic), model ID `claude-fable-5`, agentic coding session
with tool use (Claude Code / Claude Agent SDK); no extended-thinking
mode
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Related prior work: #10923 (split serialized tests into five shards),
#10925 (general-server duration manifest), #11156 (workspaces-a native
shards).
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release channel system re-smokes every published beta as
post-publish verification (`smoke_beta`)
> - The candidate-branch beta lane (#11209) added
`verify_beta_candidate` to `publish_beta`'s needs; that job is skipped
on every normal promote-mode beta
> - `smoke_beta`'s condition has no status-check function, so GitHub
attaches an implicit `success()` that evaluates the needs chain
transitively — a skipped ancestor makes it false
> - This pull request makes the condition explicit so promote-mode betas
smoke again, and pins the shape in the workflow wiring test
> - The benefit is that the post-publish beta gate actually runs instead
of silently skipping
## Linked Issues or Issue Description
**What happened?**
Beta `2026.818.0-beta.0` (run 32082007439) published successfully, but
its post-publish `smoke_beta` job was skipped. No configuration or input
asked for that: the run was a plain `channel: beta` dispatch with
`dry_run` at its default `false`, and the same expression
`!inputs.dry_run` evaluated true inside `publish_beta`'s own steps (the
Docker dispatch step ran).
**Expected behavior**
Every non-dry-run beta publish is followed by the release smoke suite
against the exact published version, as documented in `doc/RELEASING.md`
and `doc/RELEASE-CHECKLIST.md`.
**Steps to reproduce**
Dispatch `release.yml` with `channel: beta` promoting a nightly (promote
mode). `verify_beta_candidate` is skipped by design; `publish_beta` runs
through its explicit `!cancelled()` condition; `smoke_beta` then skips
because its implicit `success()` sees the skipped ancestor in the
transitive needs chain (actions/runner#2205 semantics). The beta
published on 2026-08-11 predated #11209, so this never surfaced before.
**Paperclip version or commit**
master at `43ab441f0` (workflow file, current head).
Related (not duplicates): #11209 introduced the candidate lane whose
skipped job triggers this; #11208 covers the adjacent tag-push failure
playbooks.
## What Changed
- `smoke_beta`'s condition becomes `!cancelled() &&
needs.publish_beta.result == 'success' && !inputs.dry_run` — an explicit
status-check function suppresses the implicit `success()`, and the
result check keeps the dependency on a successful publish.
- A comment above the job records why the explicit form is load-bearing.
- `scripts/__tests__/release-verify-workflow.test.mjs` pins the new
shape so the implicit form cannot silently return.
## Verification
- `node --test scripts/__tests__/release-verify-workflow.test.mjs` — 8
pass, including the new assertion.
- `release.yml` re-parsed as YAML.
- The exact skip is visible on run 32082007439 (`smoke_beta: skipped`
after `publish_beta: success`); the coverage gap for that beta was
closed manually by dispatching `release-smoke.yml` with
`paperclip_version: beta` (run 32084880767).
- Not exercised end-to-end: the corrected condition needs the next real
promote-mode beta to demonstrate; the expression change is minimal and
the semantics are the documented actions/runner behavior.
## Risks
- Low risk: condition-only change on one job plus a test. Dry runs still
skip the smoke (`!inputs.dry_run` retained). Candidate-mode betas, where
`verify_beta_candidate` actually runs, behave as before.
## Model Used
Claude Fable 5 (Claude Code)
## Pre-submission checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
> Follow-up to #11208 (merged): rebased onto master and ready for
review.
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release channels promote artifacts along canary → nightly → beta
→ stable, with the happy path being promotion of an existing build
> - When one or two targeted fixes are needed before a beta or stable,
the only options today are waiting for the next nightly or absorbing a
whole day of unrelated master changes
> - The channel model was designed with an escape hatch for exactly
this: short-lived candidate branches carrying only cherry-picked fixes
> - This pull request implements candidate-branch beta builds with full
verification, documents the stable fix path through the
soak-justification gate, and adds the release captain's checklist
> - The benefit is that a surgical fix can ship forward without either
delay or blast radius, with its provenance recorded
## Linked Issues or Issue Description
Refs #11008 — completes the fix-path half of the channel model
introduced there.
**Subsystem affected**
Release automation: `scripts/release.sh`,
`.github/workflows/release.yml`, `doc/RELEASING.md`, new
`doc/RELEASE-CHECKLIST.md`, tests.
**Problem or motivation**
Beta promotion only accepts commits that already shipped as a nightly,
and stable promotion expects a soaked beta. There is no supported way to
ship one or two cherry-picked fixes between lanes: an urgent fix must
wait for the nightly cycle or pull in every unrelated master change from
the day. The original channel design called for candidate branches to
cover this, and they were deferred from the initial implementation.
**Proposed solution**
Candidate-branch beta builds: cut `candidate/beta-<target>` from a
nightly's source commit, cherry-pick the fixes, and dispatch `channel:
beta` with the new `candidate_branch` input. Selection enforces the
naming convention, rejects heads that already shipped as a beta or
predate the candidate tooling, and records the cherry-picked commits in
the job summary. Because candidate heads never went through a canary or
nightly, publication is gated on a full `release-verify` run (promoted
nightlies keep skipping re-verification). The stable fix path
(`candidate/release-<target>` as `source_ref`) works through the
existing soak gate: the justification requirement is the deliberate,
recorded trade-off for shipping unsoaked bits, and is now documented as
such.
## What Changed
- `scripts/release.sh`: `--from-candidate` flag (beta only) waives the
shipped-a-nightly requirement while keeping the duplicate-beta guard
- `.github/workflows/release.yml`: `candidate_branch` dispatch input;
candidate mode in `select_beta` (naming validation, duplicate and
tooling-era rejection, cherry-pick recording); new
`verify_beta_candidate` job gating candidate publishes on full
verification
- `doc/RELEASING.md`: beta fix-path and stable fix-path sections
- `doc/RELEASE-CHECKLIST.md` (new): the release captain's checklist for
all four lanes as built
- Tests: dry-run fixture coverage for `--from-candidate` (waives the
nightly guard, keeps the duplicate guard, rejected outside beta) and
wiring tests for candidate validation plus the verification gate
## Verification
- `node --test` on the four affected suites: 42 pass in total (17 + 25
across the two runs), including the 5 new tests
- `bash -n` on `release.sh`; YAML parse of the workflow
- After merge: exercise the path end to end the first time a real
cherry-picked beta is needed — dispatch with a `candidate/beta-*` branch
and confirm the summary records the picks and verification runs
## Risks
- Candidate builds bypass the smoke-tested-nightly provenance by design;
the compensating controls are full verification before publish, the
post-publish beta smoke, the human `npm-beta` gate, and recorded
cherry-picks
- The stable fix path rides the existing justification mechanism rather
than adding a second bypass — one recorded escape hatch, not two
## Model Used
Claude Fable 5 (`claude-fable-5`, Anthropic) in Claude Code, with
extended thinking and full tool use. All changes model-authored under
human direction.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending — will confirm before
merge)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending — will confirm before merge)
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release subsystem's promotion lanes publish to npm, then push a
lane tag and dispatch the Docker image build at that tag
> - The first nightly of the beta-tooling merge published to npm and
then died at the tag push: GITHUB_TOKEN may not create refs pointing at
workflow-modifying commits from dispatch or scheduled runs
> - The failure was a bare `remote rejected` with no guidance, leaving
the release half-finished (npm live, no tag, no images) until an
operator reverse-engineered the recovery
> - This pull request makes every lane's tag push degrade into exact
recovery instructions in the job summary
> - The benefit is that a rare platform-permission rejection becomes a
two-minute runbook operation instead of a forensic exercise
## Linked Issues or Issue Description
Refs #11008 — the incident occurred promoting that change's own merge
commit, the first workflow-modifying commit to flow through the lanes it
introduced.
**Subsystem affected**
Release automation: `.github/workflows/release.yml`, `doc/RELEASING.md`,
workflow wiring tests.
**Problem or motivation**
Run 31445344811 published `2026.811.0-nightly.0` to npm, then failed
pushing `nightly/v2026.811.0-nightly.0`: `refusing to allow a GitHub App
to create or update workflow .github/workflows/release.yml without
workflows permission`. The tagged commit modifies workflow files, and
GITHUB_TOKEN may not create refs pointing at such commits from dispatch
or scheduled runs (push-event runs are exempt, which is why the canary
tag on the same commit succeeded). The job failed with no explanation
and the Docker dispatch never ran.
**Proposed solution**
Wrap the nightly, beta, and stable tag pushes: on rejection, write the
exact recovery commands into the job summary — create and push the tag
with maintainer credentials, dispatch `docker.yml` at the tag, and for
stable also run `create-github-release.sh` — then fail the job. Document
the cause and recovery in the failure playbooks and pin the three
recovery blocks with a wiring test.
## What Changed
- `.github/workflows/release.yml`: recovery-summary wrappers on the
nightly, beta, and stable tag-push steps
- `doc/RELEASING.md`: failure-playbook entry for the
workflows-permission rejection
- `scripts/__tests__/release-verify-workflow.test.mjs`: wiring test
asserting all three lanes carry the recovery summary
## Verification
- Wiring tests: 6 pass; YAML parse of the workflow
- The recovery commands are exactly the ones used to resolve the real
incident (tag push + `docker.yml` dispatch for
`nightly/v2026.811.0-nightly.0`)
## Risks
- Low. The happy path is unchanged (a successful push skips the
wrapper); the failure path trades a bare error for actionable output and
still fails the job, since the release state is genuinely incomplete
## Model Used
Claude Fable 5 (`claude-fable-5`, Anthropic) in Claude Code, with
extended thinking and full tool use. All changes model-authored under
human direction.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending — will confirm before
merge)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending — will confirm before merge)
- [x] I will address all Greptile and reviewer comments before
requesting merge
Split the slow workspaces-a CI lane into two Vitest native shards and keep release verification in parity.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release subsystem promotes builds along canary → nightly → beta
→ stable, and each publish job checks out the promotion's source commit
and runs that tree's release tooling
> - The first beta dispatch failed with `unexpected argument: beta`: the
selected nightly's source predated the beta channel, so its `release.sh`
did not know the argument
> - The failure was clean (argument parsing, nothing published) but
cryptic, and the same trap waits for any promotion of a source older
than its target channel's tooling
> - This pull request makes the selection jobs reject such sources with
an actionable error and documents the property
> - The benefit is that a bootstrapping or old-source promotion fails in
seconds with instructions, instead of mid-publish with a parser error
## Linked Issues or Issue Description
Refs #11008 — the guard hardens the beta promotion flow introduced
there, after its first dispatch surfaced the gap described below.
**Subsystem affected**
Release automation: `.github/workflows/release.yml`, `doc/RELEASING.md`,
workflow wiring tests.
**Problem or motivation**
Run 31444045044 (first beta dispatch) failed in `publish_beta` with
`unexpected argument: beta`. Promotions deliberately build from the
pinned source commit, which means they also run that commit's
`scripts/release.sh` — and a source that predates the target channel's
introduction cannot publish it. Nothing guards this today; the error
surfaces deep in the publish job with no explanation.
**Proposed solution**
Guard at selection time: `select_nightly` requires the source canary's
`release.sh` to know the nightly channel, and `select_beta` requires the
source nightly's `release.sh` to know the beta channel. Each guard
literally matches the channel case arm and fails closed with a clear
message naming the remedy (promote a newer source). Document the
tooling-era property in `RELEASING.md` and pin the guards with a wiring
test.
## What Changed
- `.github/workflows/release.yml`: tooling-era guards in
`select_nightly` and `select_beta`
- `doc/RELEASING.md`: documents that promotions run the source commit's
release tooling
- `scripts/__tests__/release-verify-workflow.test.mjs`: wiring test
pinning both guards
## Verification
- Wiring tests: 5 pass
- Guard expressions exercised against real commits: accepts the
beta-capable merge commit of the beta-channel change, rejects a pre-beta
commit
- YAML parse of the workflow
- After merge: the next beta dispatch selects a beta-capable nightly and
passes the guard
## Risks
- Low. Selection-time check only; the guards match the channel case arm
literally and fail closed (with the same actionable message) if that
line is ever reformatted
## Model Used
Claude Fable 5 (`claude-fable-5`, Anthropic) in Claude Code, with
extended thinking and full tool use. All changes model-authored under
human direction.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending — will confirm before
merge)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending — will confirm before merge)
- [x] I will address all Greptile and reviewer comments before
requesting merge
> Follow-up to #11006 (merged): rebased onto master and ready for
review.
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release subsystem now publishes canary (every master push),
nightly (scheduled, smoke-gated, added in #11006), and stable (manual)
> - There is still no human-approved release-candidate lane between
nightly and stable, and nothing enforces that a stable actually soaked
anywhere before shipping
> - Betas need a real approval gate, and stables need a soak policy that
is data, not prose
> - This pull request adds the beta channel: a manual promotion of a
chosen nightly behind the `npm-beta` environment gate, re-smoked after
publish, plus a stable preflight that enforces a 3-day beta soak with a
written-justification bypass
> - The benefit is a complete canary → nightly → beta → stable train
where every stable shipped as a beta first, and emergencies leave a
written trace
## Linked Issues or Issue Description
**Subsystem affected**
Release automation: `scripts/release.sh`, `scripts/release-lib.sh`,
`.github/workflows/release.yml`, `.github/workflows/docker.yml`,
`.github/workflows/release-smoke.yml`.
**Problem or motivation**
After #11006 the project has canary and nightly prerelease lanes, but no
release-candidate lane. Stable promotion has no enforced soak: any ref
can ship as stable directly. There is no approval boundary for a
broader-audience prerelease, and no structured way to record why an
emergency release skipped validation.
**Proposed solution**
Add a `beta` channel: a manual dispatch that promotes a chosen nightly's
source commit, publishes behind the `npm-beta` GitHub environment
(required reviewers are the gate), re-smokes the published beta, and
tags `beta/vX`. Enforce in the stable path that the source commit
shipped as a beta at least 3 days earlier (measured from the beta's npm
publish time), with a `skip_soak_justification` input as the recorded
emergency bypass.
**Alternatives considered**
Codifying the soak policy in docs only. Rejected: an unenforced policy
decays; the preflight makes the policy executable while the
justification input keeps the emergency path usable and auditable.
## What Changed
- `scripts/release.sh` + `scripts/release-lib.sh`: `beta` channel —
requires HEAD to carry a `nightly/v*` tag, publishes the package set as
`YYYY.MDD.P-beta.N` under dist-tag `beta`, tags
`beta/vYYYY.MDD.P-beta.N`
- `.github/workflows/release.yml`:
- `channel: beta` dispatch path: `select_beta` resolves the newest (or
an explicit `source_version`) nightly and fails loudly on selection
problems; `publish_beta` runs behind the `npm-beta` environment, pushes
the tag, and dispatches `docker.yml`; `smoke_beta` re-runs the release
smoke suite against the exact published beta version
- stable path: new `preflight_stable` job enforces the 3-day beta soak
from the beta's npm publish time; `skip_soak_justification` bypasses
with the reason echoed into the job summary; dry runs report without
blocking
- `.github/workflows/docker.yml`: `beta/v*` tags publish `:beta` on both
images, with exact version stamping
- `.github/workflows/release-smoke.yml`: `beta` added to the dispatch
choice list
- Docs: `CHANNELS.md` beta entries; `RELEASING.md` beta lane, soak gate,
and failure playbook; `RELEASE-AUTOMATION-SETUP.md` `npm-beta`
environment setup, including the warning to create the environment
before the first beta dispatch (GitHub auto-creates unprotected
environments on first reference)
- Tests: beta version-counting coverage in
`scripts/release-registry-versions.test.mjs`; beta identity and
nightly-tag guard coverage in
`scripts/__tests__/release-dry-run-notes.test.mjs`
## Verification
- `node --test` on the two touched suites: 17 pass, including the 3 new
beta tests
- `bash -n` on both shell scripts and YAML parse of all three workflows
- After merge, in order: create the `npm-beta` environment, dispatch
`channel: beta` with `dry_run: true` to preview, then a real promotion
of a published nightly through the approval gate, then a stable dry-run
against a young beta to see the soak gate report
## Risks
- If the `npm-beta` environment does not exist when the first beta
dispatch runs, GitHub creates it with no protection rules and the beta
publishes without approval. Mitigated by documentation and by creating
the environment before merge (operator step)
- Until the first beta exists, every stable dispatch requires
`skip_soak_justification`. This is deliberate — the first beta ships
immediately after this merges — but it is a behavior change to the
stable dispatch
- The soak clock reads the beta's npm publish time from the registry; a
registry outage makes the preflight fall back to requiring justification
(fail-closed)
## Model Used
Claude Fable 5 (`claude-fable-5`, Anthropic) in Claude Code, with
extended thinking and full tool use (repository exploration, local test
execution, live registry and git verification). All code, tests, and
docs in this PR were model-authored under human direction.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending — will confirm before
merge)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending — will confirm before merge)
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release subsystem's nightly lane (#11006) gates every nightly
publish on the release smoke suite, which boots the published artifact
in a Docker container
> - The suite's first CI execution failed at the health readiness check:
the harness hard-codes a 90 second budget, but a CI container
cold-installs paperclipai from npm and initializes embedded postgres
with no warm caches
> - When the timeout expired with the container still running, the
harness printed no container logs, so the failure gave no diagnostics
> - This pull request makes the readiness budget configurable, raises it
for CI, and dumps container logs on timeout
> - The benefit is that the nightly gate measures the artifact, not the
runner's cold caches, and a red smoke run is diagnosable from its logs
## Linked Issues or Issue Description
**Subsystem affected**
Release smoke testing: `scripts/docker-onboard-smoke.sh`,
`.github/workflows/release-smoke.yml`.
**Problem or motivation**
Run 31426044332 (first forced nightly after #11006) failed in
`smoke_nightly` with `server did not become ready at
http://localhost:3232/api/health` after exactly 90 seconds. The
harness's readiness window is hard-coded to 90 attempts at 1 second.
Locally that works because the npm cache is warm; in CI the container
downloads the full package set and embedded postgres first. The timeout
path also printed no container logs when the container was still
running, so there was no way to see how far boot had progressed.
**Proposed solution**
Make the readiness budget an environment variable
(`SMOKE_READY_TIMEOUT_SECONDS`, default unchanged at 90 for local use),
set it to 420 in the CI workflow, and dump the last 150 container log
lines when the readiness check times out on a still-running container.
## What Changed
- `scripts/docker-onboard-smoke.sh`: `SMOKE_READY_TIMEOUT_SECONDS` env
var (default 90) replaces the hard-coded readiness budget; timeout with
a still-running container now prints the tail of `docker logs`
- `.github/workflows/release-smoke.yml`: sets
`SMOKE_READY_TIMEOUT_SECONDS=420` for CI runs
## Verification
- `bash -n` on the harness and YAML parse of the workflow
- The real proof is the next `channel: nightly` dispatch of
`release.yml`, which re-runs this suite in CI with the new budget
## Risks
- Low. The local default is unchanged; CI runs simply wait longer before
declaring failure, and a genuinely broken artifact still fails (with
logs now)
## Model Used
Claude Fable 5 (`claude-fable-5`, Anthropic) in Claude Code, with
extended thinking and full tool use. Diagnosis from CI run logs; patch
model-authored under human direction.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending — will confirm before
merge)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending — will confirm before merge)
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The release subsystem publishes the `paperclipai` npm package set
and the Docker images on two lanes: canary on every master push, and
stable on manual promotion
> - There is no middle ground between those lanes. Users must track
every merge or wait weeks for a stable. Docker `:latest` also tracks
master, so Docker users have no stable image at all
> - A calm prerelease lane needs to exist, and it must never ship a
build that failed its checks
> - This pull request adds the nightly channel: a scheduled job that
selects the newest master commit with a green canary publish, runs the
full release smoke suite against that exact published canary, and only
then republishes it as the nightly. It also separates Docker tags by
lane, so `:latest` finally means stable
> - The benefit is that users can follow prereleases at a nightly
cadence with a smoke-tested guarantee, and Docker users get real
`:canary`, `:nightly`, and stable image tags
## Linked Issues or Issue Description
**Subsystem affected**
Release automation: `scripts/release.sh`, `scripts/release-lib.sh`,
`.github/workflows/release.yml`, `.github/workflows/docker.yml`,
`.github/workflows/release-smoke.yml`.
**Problem or motivation**
The project publishes only `canary` (every master push) and `latest`
(manual stable). Users who want prereleases without per-merge churn have
no option. Docker has a second problem: master builds overwrite
`:latest`, and CI-published stables never produced Docker images,
because tags pushed with `GITHUB_TOKEN` do not fire the `v*` tag trigger
in `docker.yml`. No stable-versioned image exists in ghcr today.
**Proposed solution**
Add a `nightly` channel. A scheduled job selects the newest
canary-tagged master commit, smoke-tests that exact published canary,
and republishes the same commit as `YYYY.MDD.P-nightly.N` under the
`nightly` dist-tag. Separate Docker tags by lane (`:canary` for master,
`:nightly` for nightly tags, `:latest` plus version tags for stable tags
only), and have the release jobs dispatch `docker.yml` at the new tag so
lane images actually build.
**Alternatives considered**
Moving the `nightly` dist-tag to the existing canary version without a
republish. Rejected: the version string would say `canary` while the
user is on nightly, which breaks at-a-glance lane identification in bug
reports and `--version` output.
## What Changed
- `scripts/release-lib.sh`: channel-parameterized
`next_prerelease_version` and `prerelease_tag_name` helpers (canary
helpers delegate to them), a `require_channel_tag_at_head` guard, and
the no-provenance retry for Sigstore transparency-log duplicates now
covers the `nightly` dist-tag as well as `canary`
- `scripts/release.sh`: new `nightly` channel. It requires HEAD to carry
a `canary/v*` tag, publishes the full public package set as
`YYYY.MDD.P-nightly.N` under dist-tag `nightly`, and tags the source
commit `nightly/vYYYY.MDD.P-nightly.N`
- `.github/workflows/release.yml`: scheduled nightly chain (09:00 UTC) —
select candidate, smoke it via `release-smoke.yml`, publish on green
under the existing `npm-canary` environment, push the tag, dispatch
`docker.yml`. New `channel` dispatch input (default `stable`, so
existing stable dispatches are unchanged) with `nightly_source_version`
and `dry_run` support for forced runs. The stable path now also
dispatches `docker.yml` at the new `v*` tag
- `.github/workflows/docker.yml`: lane tag mapping for both image jobs —
master pushes publish `:canary` and no longer move `:latest`;
`nightly/v*` tags publish `:nightly`; only stable `v*` tags publish
`:latest` and the versioned tags. New `workflow_dispatch` trigger for
the release-job dispatches. Build-version stamping uses the exact
nightly version on nightly tag builds
- `.github/workflows/release-smoke.yml`: `nightly` added to the dispatch
choice list
- `doc/CHANNELS.md` (new): user-facing guide to the channels
- `doc/RELEASING.md`: nightly lane documentation, Docker tag mapping
table, and a nightly failure playbook
- `doc/RELEASE-AUTOMATION-SETUP.md`: note that nightly reuses
`npm-canary` and needs no npm trusted-publisher changes
- Tests: channel-parameterized version helper coverage in
`scripts/release-registry-versions.test.mjs`, and nightly flow coverage
(publish identity, notes not required, canary-tag guard) in
`scripts/__tests__/release-dry-run-notes.test.mjs`
## Verification
- `node --test` on the release script suites: 68 pass, including 6 new
tests. The only failure, `acpx-patch-packaging.test.mjs`, needs
installed `node_modules` and fails identically on a pristine checkout of
master in the same environment
- `bash -n` on both shell scripts and YAML parse of all three workflows
- Live fail-path check: `./scripts/release.sh nightly --print-version`
from a master tip with no canary tag fails with `HEAD has no canary/v*
tag`
- Live success-path check: the same command from the
`canary/v2026.806.0-canary.7` commit prints `2026.806.0-nightly.0`
- Live selection check: the candidate-selection shell logic run against
the real repository selects the commit of `canary/v2026.806.0-canary.7`,
which matches the current npm `canary` dist-tag exactly
- After merge: dispatch `release.yml` with `channel: nightly` and
`dry_run: true` to preview, then a real forced run to validate end to
end before the first scheduled run
## Risks
- Docker `:latest` changes meaning from "latest master build" to "latest
stable release". This is deliberate and will be announced. Users who
want the old behavior pull `:canary`. Until the first stable release
after this change, `:latest` stays at its current (master-built) image
- The nightly is a rebuild of the same source commit, not the
byte-identical canary artifact that was smoked. The lockfile pins
dependencies, and the publish path's registry-visibility and
clean-prefix install gates still run on the nightly artifacts
- All npm publishing must stay inside `release.yml` because npm trusted
publishing pins that workflow file per package. The nightly jobs were
added to `release.yml` for exactly that reason; this constraint is now
documented in `RELEASING.md`
- The stable-lane Docker dispatch fails gracefully (a warning with
manual instructions) when the source ref predates `docker.yml`'s
`workflow_dispatch` trigger
## Model Used
Claude Fable 5 (`claude-fable-5`, Anthropic) in Claude Code, with
extended thinking and full tool use (repository exploration, local test
execution, live registry and git verification). All code, tests, and
docs in this PR were model-authored under human direction.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending — will confirm before
merge)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending — will confirm before merge)
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The PR workflow runs the server vitest suite across sharded runners
because the suite is pinned to one worker.
> - In successful PR run 30930345729 (2026-08-04), shard `server (3/4)`
took 311 seconds of wall time and was the slowest check in the run.
> - The suite has grown to about 946 seconds of serial vitest time, but
the duration manifest was last sampled on 2026-08-01 at about 882
seconds.
> - This pull request refreshes the per-suite duration manifest from
that run's logs and splits the lane into five shards.
> - The benefit is a shorter PR critical path: each shard carries about
196 seconds of suite time, level with the other lanes.
## Linked Issues or Issue Description
Refs #10663 (previous split of this lane into four shards).
Related: #10923 splits the separate serialized-suites lane into five
shards. Both PRs touch `.github/workflows/pr.yml` in different matrix
blocks; whichever merges second needs a trivial rebase.
**What existing behavior does this improve?**
The `general-server` vitest lane runs in four shards with a duration
manifest sampled on 2026-08-01.
**Current behavior**
In PR run 30930345729, shard 3/4 ran for 311 seconds (273 seconds in the
test step) and was the longest check in the run. The suite now totals
about 946 seconds of serial vitest time.
**Proposed behavior**
Run the same suite set in five shards, balanced with a per-suite
duration manifest refreshed from that run's shard logs (279 suites
measured by diffing consecutive completion timestamps).
**Reason and benefit**
The refreshed LPT partition balances at about 196 seconds of suite time
per shard (about 240 seconds per job), level with the other PR lanes. No
test coverage is lost.
**Breaking changes**
None. The change only alters the CI partition size and the duration
manifest.
## What Changed
- Bump the `general-server` shard matrix in `.github/workflows/pr.yml`
from four to five shards.
- Refresh `scripts/general-server-shard-durations.json` from the
2026-08-04 run's shard logs.
- Update `SHARD_COUNT` in
`scripts/__tests__/run-vitest-stable-shard.test.mjs` to five.
## Verification
- `node --test scripts/__tests__/run-vitest-stable-shard.test.mjs` — 9/9
pass, including the complete non-overlapping partition proof and the
duration-balance check.
- `node --test scripts/__tests__/release-verify-workflow.test.mjs` — 2/2
pass.
- `node --test scripts/__tests__/e2e-shard.test.mjs` — 7/7 pass.
- A 5-way dry-run partition covers all suites exactly once with equal
projected weights.
## Risks
- Low risk. The change only alters CI partition size and duration
weights; the suite set is unchanged.
- One more runner is used per PR run for this lane.
- Stale duration weights degrade gracefully: suites missing from the
manifest get the median weight.
## Model Used
- Claude (Anthropic), Claude Code CLI, model ID `claude-fable-5`,
extended thinking with tool use enabled.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
(workflow comments explain the new shard math)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Claude <claude@paperclip.ing>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip uses CI to keep control-plane changes safe and mergeable.
> - The PR workflow splits serialized server tests across isolated
runners.
> - A recent successful run spent 305 seconds in serialized shard 2/4.
> - That job was the slowest check in the run.
> - The four shards reported about 739 seconds of Vitest suite time.
> - This pull request adds a fifth serialized shard and keeps release
verification aligned.
> - The benefit is a shorter PR critical path with no loss of test
coverage.
## Linked Issues or Issue Description
**What existing behavior does this improve?**
The PR and release verification workflows run serialized server tests in
four shards.
**Current behavior**
Successful PR run 30876682788 spent 305 seconds in `Verify serialized
server suites (2/4)`. The test step used 256 seconds and made this job
the slowest check.
**Proposed behavior**
Run the same serialized suite set in five complete and non-overlapping
shards.
**Reason and benefit**
The measured suites reported about 739 seconds of total Vitest time.
Five runners reduce the expected average suite time from about 185
seconds to about 148 seconds before setup overhead.
**Breaking changes**
None. The change only alters CI partition size.
## What Changed
- Split serialized server tests into five shards in the PR workflow.
- Apply the same five-shard layout to release verification.
- Add a partition test that proves complete and non-overlapping
serialized coverage.
- Update release workflow coverage tests for five shards.
## Verification
- `node --test scripts/__tests__/run-vitest-stable-shard.test.mjs
scripts/__tests__/release-verify-workflow.test.mjs`
- `git diff --check`
## Risks
- Low risk. CI uses one additional runner for the serialized lane.
- Round-robin partition weights can still vary as suite timings change.
> This change does not overlap with planned core work in `ROADMAP.md`.
Related PR #10663 optimized the separate general-server lane.
## Model Used
- OpenAI Codex, GPT-5, agentic coding with reasoning, tool use, and code
execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [ ] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Devin Foley <139239+devinfoley@users.noreply.github.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Pull request CI must give contributors fast and stable feedback.
> - The `general-server` Vitest lane runs many single-worker server
suites.
> - A recent completed PR run showed this lane as the slowest completed
check.
> - Three shards still left one runner with the largest share of work.
> - This pull request splits that lane into four duration-balanced
shards.
> - The benefit is a shorter critical path for the same server test
coverage.
## Linked Issues or Issue Description
No public GitHub issue exists for this CI maintenance change.
**Pre-submission checklist**
- I confirmed this improves existing behavior. It does not add a new
command, endpoint, or concept.
- I searched open public issues and pull requests for related CI
sharding work.
**What existing behavior does this improve?**
The pull request workflow's `general-server` Vitest lane.
**Subsystem affected**
Cross-cutting. This affects GitHub Actions CI and the Vitest shard
duration manifest.
**Current behavior**
The `general-server` lane uses three shards. The server suites now total
about 880 seconds of serial Vitest wall time. The slowest shard was
about 313 seconds in the measured run.
**Proposed behavior**
The `general-server` lane uses four shards. Each shard receives about
220 seconds of predicted suite weight from the refreshed duration
manifest.
**Reason and benefit**
The slowest PR check controls how soon a reviewer can trust the PR. Four
balanced shards reduce the slowest `general-server` shard while keeping
the same suite selection rules.
**Breaking changes**
None. This only changes CI partitioning and duration data for existing
test suites.
**Additional context**
Related public searches found no exact open issue or pull request for
this `general-server` sharding change.
## What Changed
- Split the `general-server` CI matrix from three shards to four shards.
- Refreshed `scripts/general-server-shard-durations.json` with wall-time
weights from a recent completed PR run.
## Verification
- `node --test scripts/__tests__/run-vitest-stable-shard.test.mjs`
- `git diff --check origin/master...HEAD`
- Dry-ran the four `general-server` shards locally during
implementation. The partition covers 300 unique suites with about 220.56
seconds of predicted weight per shard.
- Ran a local sensitive-data scan before push. It found only test
filenames that contain words such as `secret` or `token`, not credential
values.
## Risks
Low risk. The main risk is that the duration manifest becomes stale as
suite costs move. Missing suites fall back to the median weight, so the
lane still runs if the manifest is incomplete.
## Model Used
OpenAI Codex, GPT-5, with tool use and local command execution.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip uses pull request CI to test changes before merge.
> - The e2e PR lane runs Playwright specs in a shard matrix.
> - Each shard builds a list of spec files for its matrix entry.
> - The workflow passed that list after a literal `--` separator.
> - Playwright did not receive the list as file filters.
> - This pull request removes the separator and adds a guard test.
> - The benefit is that each e2e shard runs only its assigned specs.
## Linked Issues or Issue Description
Refs #10629.
**What happened?**
The e2e shard step used `pnpm run test:e2e -- $specs`. The shard spec
list was not applied as Playwright file filters.
**Expected behavior**
Each e2e shard should pass only its selected specs to Playwright.
**Steps to reproduce**
1. Inspect `.github/workflows/pr.yml` at the merge commit for #10629.
2. Find the `e2e_shards` command that invokes `pnpm run test:e2e`.
3. See the literal `--` before `$specs`.
**Paperclip version or commit**
`86767951`
**Deployment mode**
GitHub Actions PR CI.
## What Changed
- Removed the literal `--` from the e2e shard `pnpm run test:e2e $specs`
invocation.
- Added a regression test that checks the workflow passes `$specs`
without that separator.
## Verification
- `node --test scripts/__tests__/e2e-shard.test.mjs`
## Risks
Low risk. This changes one CI command and one workflow guard test. The
main risk is shell argument handling in the workflow, and the guard now
covers the expected command shape.
## Model Used
OpenAI GPT-5 through Codex. The run used shell and GitHub CLI tool
access. The runtime did not expose a context window size.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The pull request workflow protects changes with a Playwright e2e
lane.
> - That lane already uses a weighted file partition so slow specs do
not cluster by test count.
> - Recent green PR runs showed the two e2e shard jobs were slower than
the next slow required lane.
> - The largest spec is indivisible, so a third shard lets that spec run
alone and lets the rest split by duration.
> - This pull request changes only the PR e2e shard matrix and the guard
test.
> - The benefit is a shorter expected PR critical path while the
required `e2e` aggregate check name stays stable.
## Linked Issues or Issue Description
Refs #9923
**What existing behavior does this improve?**
The `pull_request` workflow Playwright e2e lane.
**Subsystem affected**
Cross-cutting: GitHub Actions CI and test scripts.
**Current behavior**
The PR workflow runs the weighted Playwright e2e partition across two
jobs. Recent green runs showed those jobs as the slowest required
checks.
**Proposed behavior**
The PR workflow runs the same e2e spec set across three weighted jobs.
The aggregate required check stays named `e2e`.
**Reason and benefit**
The third shard lets the slow smoke-lab spec run alone while the rest of
the catalog stays balanced. This should shorten the PR critical path.
The win is bounded by fixed per-job setup time.
**Breaking changes**
None. The required aggregate check contract is preserved.
## What Changed
- Change the PR e2e shard matrix from two entries to three entries.
- Update the shard guard test to expect three shards.
- Floor the balance bound at the largest single spec weight.
- Assert that the workflow does not define more shard indexes than
`SHARD_COUNT`.
## Verification
- `node --test ./scripts/__tests__/e2e-shard.test.mjs` passes with 6
tests.
- The recorded-weight partition is complete and non-overlapping: 168.0s,
116.5s, and 114.4s.
- I checked `ROADMAP.md` and found no overlapping roadmap-level core
feature.
- I searched public GitHub PRs and issues for related e2e shard work. I
found related PR #9923 and no open duplicate for this branch or change.
## Risks
- This adds one extra GitHub Actions runner to the PR e2e lane.
- The wall-clock win is bounded by fixed per-job setup.
- Behavior risk is low because the aggregate required check remains
named `e2e`.
## Model Used
OpenAI Codex, GPT-5, tool-enabled coding agent in this repository. The
runtime did not expose the context-window size.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Cody <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Published container images are the deployable unit for self-hosted
and managed instances, so merge-to-image latency bounds every deploy
iteration
> - The docker workflow configures BuildKit caching, but builds still
ran ~12+ minutes essentially cold
> - Two causes: the most expensive layer (four CLI toolchains + apt) is
ordered after the always-changing app copy so it can never cache, and
the type=gha cache's 10GB repo cap means the two multi-arch mode=max
jobs evict each other
> - This pull request reorders the tool layer above the app copy (with a
weekly epoch so @latest tools keep advancing) and switches both jobs to
registry-backed cache in ghcr
> - The benefit is that warm builds shrink to roughly the app build +
push, targeting the sub-5-minute range together with the amd64-only
cloud variant
## Linked Issues or Issue Description
No existing public issue — inline description following the feature
request template:
**Subsystem affected**
CI / release publishing (docker workflow, Dockerfile)
**Problem or motivation**
Despite `cache-from/cache-to` being configured, image builds run
effectively cold: (1) the production stage installs four CLI toolchains
+ apt packages *after* `COPY --from=build /app /app`, and since the app
copy changes every commit, that most-expensive layer rebuilds every
build, per arch; (2) the `type=gha` BuildKit cache is capped at 10GB per
repository, and two multi-arch `mode=max` jobs overflow and evict each
other's entries.
**Proposed solution**
Order the tool/OS layer before the app copy (it references nothing from
`/app`), refresh it weekly via a `CLI_TOOLS_CACHE_EPOCH` build arg so
the `@latest` tools don't freeze in the cache, and move both jobs to
registry-backed BuildKit cache (`:buildcache` / `:buildcache-cloud` refs
in ghcr, no size cap, separate refs so the parallel jobs don't clobber
each other).
**Alternatives considered**
Pinning CLI tool versions instead of the weekly epoch — more
deterministic, but adds a version-bump chore; the weekly epoch preserves
current freshness semantics with bounded staleness. Keeping type=gha
with `mode=min` — smaller cache but loses intermediate-stage reuse,
which is where most of the win is.
**Roadmap alignment**
Not on ROADMAP.md; CI/publishing speed improvement only.
## What Changed
- `Dockerfile`: the production stage's tool/OS `RUN` (npm --global CLIs,
apt, `/paperclip` setup) moves above `COPY --from=build /app /app`; new
`CLI_TOOLS_CACHE_EPOCH` arg consumed by that layer. The `cloud` stage is
unaffected — it only layers plugin dists on top of the finished
production stage.
- `.github/workflows/docker.yml`: both jobs stamp the ISO week into
`CLI_TOOLS_CACHE_EPOCH`, and both switch `cache-from/cache-to` from
`type=gha` to `type=registry` with per-job refs.
- Includes the one-line amd64-only cloud-variant commit from #10570 so
the two PRs can't conflict; if #10570 merges first, this PR rebases down
to a single commit automatically.
## Verification
- Image content is unchanged by layer reordering: the moved `RUN`
references nothing from `/app`, and Docker layer ordering only affects
caching, not the final filesystem (tool installs and app copy touch
disjoint paths).
- The cache ref is written only by this workflow — `docker.yml` runs on
master/tag pushes, never on PRs — so the workflow's existing "no shared
caches into build inputs" supply-chain stance is unchanged (BuildKit
layer cache was already accepted via type=gha; the registry backend has
the same writer trust).
- Runtime proof lands with the first two master builds after merge: the
first warms the cache, the second should show the tool layer and
deps/build stages as CACHED in the build log, with wall clock dropping
accordingly. I'll be watching those as part of managed-deploy work.
## Risks
- Low. Worst case the registry cache misses (cold-build behavior, same
as today). The weekly epoch means CLI tools update at most a week late
inside images; a release built mid-week ships the tools from that week's
first build. Cache refs add two small artifacts to ghcr.
## Model Used
Claude Fable 5 (`claude-fable-5`, extended thinking, via Claude Code
with tool use and code execution).
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (no test-affecting changes)
- [x] I have added or updated tests where applicable (n/a — build config
and layer ordering only)
- [x] I have updated relevant documentation to reflect my changes
(in-file comments document both mechanisms)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Published container images are how both self-hosted users and
managed-deployment hosts run it
> - The repository publishes two variants: the self-hosted image and a
cloud variant for managed deployments
> - The cloud variant was built for amd64+arm64, but its only consumers
are managed-deployment hosts, which run amd64
> - The QEMU-emulated arm64 half dominates the build's wall clock,
delaying every merge-to-deployable-image cycle
> - This pull request drops arm64 from the cloud variant only, keeping
the self-hosted image multi-arch
> - The benefit is roughly halving the time from merge to a deployable
cloud image, with no change for any actual consumer
## Linked Issues or Issue Description
No existing public issue — inline description following the feature
request template:
**Subsystem affected**
CI / release publishing (docker workflow)
**Problem or motivation**
The cloud image variant builds for `linux/amd64,linux/arm64`, but the
arm64 half runs under QEMU emulation and dominates the job's wall clock
— while no consumer of the cloud variant runs arm64 (managed-deployment
hosts are amd64). Every deploy iteration pays ~double the necessary
build time.
**Proposed solution**
Build the cloud variant amd64-only. The self-hosted image keeps
`amd64+arm64` so ARM users (Apple Silicon, ARM servers) are unaffected.
**Alternatives considered**
Keeping multi-arch but building arm64 on native arm64 runners with a
manifest merge — faster than QEMU and worth doing for the self-hosted
image if its build time becomes a pain point, but unnecessary complexity
for a variant with no arm64 consumers.
**Roadmap alignment**
Not on ROADMAP.md; CI/publishing speed improvement only.
## What Changed
- `.github/workflows/docker.yml`: the `build-and-push-cloud` job's
`platforms` is now `linux/amd64` (with a comment explaining why). The
self-hosted `build-and-push` job is untouched.
## Verification
- Build-config-only change; the workflow runs on merge to master. The
published `-cloud` manifest will be amd64-only, which its consumers
already pull.
- No test changes: nothing at runtime differs on any platform that
actually runs the image.
## Risks
- Low. If an arm64 consumer of the cloud variant ever appears (e.g.
local `docker run` on Apple Silicon for debugging), it would fall back
to emulation on the consumer's machine or need this reverted — a
one-line change. The self-hosted image's platform matrix is unchanged.
## Model Used
Claude Fable 5 (`claude-fable-5`, extended thinking, via Claude Code
with tool use and code execution).
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (no test-affecting changes)
- [x] I have added or updated tests where applicable (n/a — CI platform
matrix only)
- [x] I have updated relevant documentation to reflect my changes
(in-workflow comment documents the rationale)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Instances commonly run from published container images, and
operators need to observe which build a container actually serves
> - `/api/health` now reports the running build commit, and server-info
already falls back to `PAPERCLIP_BUILD_COMMIT` when git is unavailable
> - But published images carry no `.git` and never received
`PAPERCLIP_BUILD_COMMIT`, so containers report `commit: null` — verified
live against a current image
> - That leaves the new deployment-verification field inert exactly
where it matters most: containerized deploys
> - This pull request bakes the exact build commit into both image
variants at build time, mirroring how `PAPERCLIP_BUILD_VERSION` is
already stamped
> - The benefit is that containers report their true commit on
`/api/health`, so deploy tooling can verify a rollout actually shipped
## Linked Issues or Issue Description
Companion to #10563 (which exposed the `commit` field on `/api/health`).
Inline description following the bug report template:
**What happened?**
A container from a published image responds to `GET /api/health` with
`"commit": null`. The image has no `.git` directory and the
`PAPERCLIP_BUILD_COMMIT` fallback that `server-info` supports is never
provided at build time, so git metadata resolves as unavailable.
**Expected behavior**
A container reports the commit it was built from, the same way it
already reports its build version via the baked
`PAPERCLIP_BUILD_VERSION`.
**Steps to reproduce**
Run any published image (e.g.
`ghcr.io/paperclipai/paperclip:sha-c4f6264-cloud`) and `curl
/api/health` — `commit` is `null` even though the build commit is known
at image-build time.
**Paperclip version or commit**
`sha-c4f6264-cloud` (first image containing #10563).
## What Changed
- `Dockerfile`: new `PAPERCLIP_BUILD_COMMIT` build arg, exported as an
ENV in the production stage (the `cloud` stage inherits it), directly
parallel to `PAPERCLIP_BUILD_VERSION`. Empty for local `docker build`,
which keeps the normal fallbacks.
- `.github/workflows/docker.yml`: both build jobs pass
`PAPERCLIP_BUILD_COMMIT=${{ github.sha }}`.
## Verification
- Reviewed the plumbing end-to-end: `build-commit.ts` reads
`PAPERCLIP_BUILD_COMMIT` (validated as a full SHA), `server-info.ts`
`readGitInfo` falls back to it when the git CLI fails, producing
`available: true, fullSha` — which `/api/health` surfaces as `commit`.
- Verified live that a current published image reports `commit: null`;
this change repairs that on the next build. Post-merge, the first master
image should report its commit — I'll be verifying that as part of
managed-deploy validation.
- No test changes: the fallback path is already covered by existing
server-info tests; this PR only supplies the env at image build.
## Risks
- Low. Two build-time stamps; no runtime code changes. A wrong SHA would
only mislabel the build (same failure mode `PAPERCLIP_BUILD_VERSION`
already carries), and `${{ github.sha }}` is the exact commit the
workflow builds.
## Model Used
Claude Fable 5 (`claude-fable-5`, extended thinking, via Claude Code
with tool use and code execution); diagnosis included live probes of a
running container's `/api/health`.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (no test-affecting changes;
server suites unaffected)
- [x] I have added or updated tests where applicable (n/a — build-time
stamps only)
- [x] I have updated relevant documentation to reflect my changes
(Dockerfile comments document the arg)
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The CI workflow is responsible for producing and publishing the
Docker images that power Paperclip deployments
> - The cloud image publish was previously coupled to the stock publish
job, so a failure or delay in one path could gate the other
> - That coupling makes the release pipeline less resilient than it
needs to be
> - This pull request gives the cloud publish its own top-level job so
both publishes can run in parallel without a `needs:` dependency
> - The benefit is better failure isolation and less wasted time when
one publish path is slow or broken
## Linked Issues or Issue Description
No public GitHub issue was found for this change.
Problem statement:
- The cloud image publish was implemented as trailing work inside the
stock publish job.
- That setup meant the cloud publish could be delayed or skipped if the
stock job failed early.
- The desired behavior is for the cloud publish to run independently so
a failure in one publish path does not gate the other.
Proposed solution:
- Split the cloud publish into its own top-level workflow job.
- Keep the same cloud-specific build settings and cache behavior.
- Preserve the existing top-level concurrency behavior.
Alternatives considered:
- Keeping both publishes in one job with conditionals or later steps.
Rejected because it still couples success and runtime between the two
publish paths.
## What Changed
- Split the cloud image publish into a separate top-level Docker
workflow job.
- Removed the dependency coupling so the cloud job does not need the
stock job.
- Expanded the drift-guard test to assert the two-job structure and the
absence of `needs:` on the cloud job.
## Verification
- The workflow YAML was parsed successfully and confirmed to contain two
jobs: `build-and-push` and `build-and-push-cloud`.
- The cloud job was confirmed to have no `needs:` entry.
- The drift-guard assertions were reproduced in a dependency-free
harness and passed.
- PR #10408 completed GitHub Actions with all required checks green,
including the e2e shards.
- Greptile review completed at 5/5 with no unresolved comments.
- No documentation files changed because this is a workflow/test-only
change.
## Risks
- The workflow now duplicates the prep steps across two runners, so any
shared setup change must be kept in sync between both jobs.
- The new job increases workflow surface area slightly, which can make
future maintenance more verbose.
- Overall risk is low because the change is limited to CI orchestration
and test coverage.
## Model Used
OpenAI Codex (GPT-5, tool-using code assistant)
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The repository uses GitHub Actions workflows to build and publish
Docker images
> - A workflow-level concurrency policy controls whether newer pushes
cancel in-flight builds
> - The current job-level setting can cancel a running image build
before it finishes publishing
> - That leaves the Docker image pipeline brittle when a new push
arrives during an active publish
> - This pull request moves concurrency to the workflow level and
disables cancel-in-progress so running builds finish
> - The benefit is that only pending work is superseded, while a build
already publishing is allowed to complete
## Linked Issues or Issue Description
No public GitHub issue exists for this change. This PR addresses the
Docker workflow concurrency behavior directly: it ensures in-flight
image builds are not cancelled by newer pushes, while still serializing
builds per ref.
## What Changed
- Moved the Docker workflow concurrency block from the job level to the
workflow level.
- Set `cancel-in-progress: false` so an active build can finish
publishing.
- Added a drift-guard test that parses `.github/workflows/docker.yml`
and asserts the workflow-level concurrency policy remains `false`.
## Verification
- `pnpm --filter @paperclipai/server exec vitest run
src/__tests__/cloud-image-bundled-plugins.test.ts`
- `python3 -c "import yaml;
yaml.safe_load(open('.github/workflows/docker.yml'))"`
- Verified the fetched remote branch contains a single commit on top of
`origin/master`.
- Searched GitHub for duplicate or related PRs and issues; none found.
- Checked `ROADMAP.md` and did not find overlapping planned core work.
## Risks
- Low risk: the change is limited to workflow concurrency behavior and a
targeted test assertion.
- If the workflow concurrency key is changed later, the drift-guard test
will fail and require an update.
## Model Used
OpenAI Codex (GPT-5, tool use; context window not surfaced in this
environment)
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` /
`Closes: #` / `Refs: #` OR (b) described the issue in-PR following the
relevant issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work; it ships as a Docker image that self-hosters and managed
deployments run.
> - The server resolves its own version at runtime in
`server/src/version.ts` (`resolveServerVersion()`), which feeds
analytics and the server debug panel.
> - That resolver derives the real version from `git describe`, and
falls back to `server/package.json`'s `version` when git isn't
available.
> - But `server/package.json`'s version is a static placeholder — CI
only stamps the real CalVer at publish, so in source it is never the
real version (currently `0.3.1`).
> - A Docker image has no `.git` (it's dockerignored), so `git describe`
can't run inside it. Every image therefore falls back to the placeholder
and reports `0.3.1` in analytics and the debug panel, regardless of
which commit it was built from.
> - This PR computes the real version once on the CI build runner (where
`.git` and tags exist), bakes it into the image, and has
`resolveServerVersion()` prefer that stamp when `git describe` is
unavailable.
> - The benefit: self-hosted and cloud images report their true version
instead of a misleading placeholder, with no change to dev checkouts,
`git describe`-based resolution, or local `docker build`.
## Linked Issues or Issue Description
No public issue exists — describing the bug inline (per the bug report
template).
**What happened?**
Docker images built from `master` (and release tags) report the server
version as the `0.3.1` placeholder in analytics and the server debug
panel, instead of the real version of the commit the image was built
from.
**Expected behavior**
An image reports the real version of its build commit (e.g.
`2026.722.0+51.git.<sha>`), so operators can tell which build is
running.
**Steps to reproduce**
1. Build the server Docker image from any `master` commit (the `Docker`
workflow, `production` target).
2. Run the image and open the server debug panel (or inspect the version
reported to analytics).
3. Observe the version is `0.3.1` rather than the commit's real version.
**Root cause**
`resolveServerVersion()` derives the real version from `git describe`,
but the image has no `.git` (dockerignored), so it falls back to
`server/package.json`'s `version` — a static placeholder CI only
replaces with the real CalVer at publish time. Nothing bakes the real
version into the image.
**Paperclip version or commit:** reproduces on `master` (`4c55f0d8`) and
any published image.
**Deployment mode:** self-hosted and managed (both the `production` and
`-cloud` images).
**Installation method:** Docker image (`ghcr.io/paperclipai/paperclip`).
**Related PRs (dedup search):** #9103 (merged — added the `git
describe`-based source-install resolution this builds on) and #9637
(closed). Neither bakes a version into the image; this PR closes that
gap. No duplicate found.
## What Changed
- **`.github/workflows/docker.yml`** — checkout with full history + tags
(`fetch-depth: 0`), and a new `Compute build version` step that runs
`git describe --tags --match 'v*' --long --dirty` on the pristine runner
checkout. The result is passed as a `PAPERCLIP_BUILD_VERSION` build-arg
to both the `production` and `-cloud` image builds.
- **`Dockerfile`** — the `production` stage takes an `ARG
PAPERCLIP_BUILD_VERSION` (default empty) and bakes it into the runtime
`ENV`; the `cloud` stage inherits it via `FROM production`.
- **`server/src/build-version.ts`** (new) — `readBuildVersion()` /
`parseBuildVersion()`, mirroring `build-commit.ts`: reads
`PAPERCLIP_BUILD_VERSION` (or a `.paperclip-build-version` file) as a
single-token stamp.
- **`server/src/version.ts`** — `resolveServerVersion()` prefers the
baked build version when `git describe` is unavailable, parsing it with
the same rules as a live checkout (`parseGitDescribeVersion`), and
falling through to the existing `build-commit` stamp and package version
when unset. A live checkout's `git describe` still wins over any stamp.
- Tests for the new behavior and the precedence.
## Verification
- `pnpm --filter @paperclipai/plugin-sdk ensure-build-deps && tsc
--noEmit` in `server/` — clean.
- `vitest run server/src/__tests__/version.test.ts
server/src/__tests__/build-version.test.ts` — **23 tests pass**,
covering: stamped version used when git describe fails, stamp parsed to
real CalVer, stamp preferred over the build-commit fallback, on-tag
stamp collapses to the release version, a pre-resolved stamp used
verbatim, and a live git describe still winning over a stamp.
- `git describe --tags --match 'v*' --long` for this commit →
`v2026.722.0-51-g<sha>`, which `resolveServerVersion()` reports as
`2026.722.0+51.git.<sha>` — no longer `0.3.1`.
- Not run locally: the full multi-arch image build (CI-only). The
workflow change is verified by inspection; the version is computed on
the pristine checkout before any lockfile refresh, so it carries no
spurious `-dirty`.
## Risks
Low. Additive and image-only:
- No runtime behavior changes for dev checkouts (git describe still
primary and wins over any stamp) or for local `docker build` (empty arg
→ server keeps its existing fallbacks).
- Not a breaking change; no schema or API surface. The stamp is
informational (version reporting only).
- `fetch-depth: 0` makes the release-image checkout fetch full
history/tags — a modest cost on a workflow that already runs at release
cadence with a 60-minute budget.
- Rollback: revert the commit; images simply return to reporting the
placeholder.
## Model Used
Claude Opus 4.8 (`claude-opus-4-8`, 1M-context variant), extended
thinking, with tool use / code execution — agentic edits, `tsc` +
`vitest` runs, and a `git describe` resolution check.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work (bugfix, not core feature work)
- [x] I have searched GitHub for duplicate or related PRs and linked
them above (#9103, #9637 — related, not duplicates)
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (`fix/build-version-stamp`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes (no
user-facing docs affected; behavior is documented inline in `version.ts`
/ `build-version.ts` and the workflow/Dockerfile)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Managed (cloud-hosted) deployments configure instances through
`PAPERCLIP_MANAGED_CONFIG`, including a `plugins.autoInstall` key list
that the boot-time installer resolves against the bundled plugin catalog
> - The installer requires each bundled plugin's `dist/manifest.js`
(`server/src/services/bundled-plugins.ts`), but the published image only
ships the sandbox providers' *source* — they are intentionally excluded
from the pnpm workspace, and the Dockerfile never builds them
> - Every managed auto-install therefore logs `bundled plugin bundle not
present; skipping auto-install` and no sandbox provider can be
provisioned through managed config
> - Baking built plugins into the single published image would fix it
but makes every self-hosted pull carry the providers' `node_modules` for
a managed-only mechanism
> - This pull request adds a `cloud` Dockerfile target extending
`production` with built bundled plugins — parameterized by build arg and
currently just `daytona` — published alongside the default image with a
`-cloud` tag suffix
> - The benefit is working plugin auto-provisioning for managed
deployments while the self-hosted image stays byte-identical and the
cloud variant only carries what is actually deployed
## Linked Issues or Issue Description
Fixes#10158 (filed for this problem; no prior issue existed — searched
for duplicate/related PRs and issues around bundled plugins, docker
image variants, and auto-install). Summary: **What happened:** on a
managed instance with `plugins.autoInstall: ["daytona"]` delivered via
`PAPERCLIP_MANAGED_CONFIG`, boot logs `bundled plugin bundle not
present; skipping auto-install` with `pluginPath:
/app/packages/plugins/sandbox-providers/daytona`, and the plugin is
never installed. **Expected:** the advertised bundled-catalog keys are
installable from the published image. **Why:** the image ships plugin
source without `dist/` — nothing in the Dockerfile builds the
workspace-excluded sandbox providers.
## What Changed
- `Dockerfile`: new `cloud-plugins` stage (based on `build`, so
devDependencies are available for `tsc`) that installs and builds each
provider named in the `CLOUD_BUNDLED_PLUGINS` build arg standalone
(`pnpm install --ignore-workspace --no-lockfile && pnpm build`, exactly
as the providers' READMEs prescribe), asserting `dist/manifest.js`
exists per plugin and failing loudly on unknown names; new `cloud` stage
= `production` + the built plugin tree. The arg defaults to `daytona` —
the only provider managed deployments auto-install today; every entry
adds its `node_modules` to the image, so the list grows only with actual
need (a one-line workflow change).
- `.github/workflows/docker.yml`: the existing build step is pinned to
`target: production` (without this, the new trailing stage would
silently become the default build target — this pin is what keeps the
self-hosted image identical); new metadata + build-push steps publish
the `cloud` target (with `CLOUD_BUNDLED_PLUGINS=daytona`) under the same
tag set with a `-cloud` suffix (`sha-<short>-cloud`, `latest-cloud`,
`<version>-cloud`), same schema labels, reusing the GHA layer cache
## Verification
- All seven sandbox providers build standalone from a clean checkout
with the exact commands the new stage runs, each producing
`dist/manifest.js` — so the current `daytona` default works and future
list additions are known-good
- The stage's shell loop was dry-run against the checkout (directory
existence + per-plugin assertion logic)
- Workflow YAML lints clean
- **Not run:** a full multi-arch `docker build` (no local docker
daemon). The `cloud` stage is additive and the default target is pinned,
so the risk is contained to the new build step; the first master build
after merge proves it end-to-end
## Risks
- Self-hosted behavior: unchanged. The default image build is pinned to
the `production` target, which produces the same layers as before this
change; the `cloud` stages run only for the new build step.
- The plugin installs in the `cloud-plugins` stage use `--no-lockfile`
(the providers are workspace-excluded and lockfile-less by design), so
plugin dependency resolution is not pinned at image-build time. This
mirrors the existing Plugins-page install path, which resolves from npm
at install time.
- CI cost: one additional build-push per master push. It reuses the
layer cache from the production build, so the marginal work is the
single plugin's build layers.
- An unknown name in `CLOUD_BUNDLED_PLUGINS`, or a provider that stops
producing `dist/manifest.js`, fails the cloud build loudly rather than
publishing a broken variant.
## Model Used
Claude (Anthropic), model ID `claude-fable-5[1m]` via Claude Code CLI —
extended thinking and tool use (code edits, standalone plugin build
verification, workflow lint).
## Checklist
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] Self-hosted behavior unchanged (default build target pinned to
`production`)
- [x] One clear change: publish a cloud image variant with built bundled
plugins
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The Docker image publish workflow (`.github/workflows/docker.yml`)
builds and pushes the multi-arch `ghcr.io` image on every master push,
so users pulling the container get the latest code
> - The two newest master runs of that workflow failed, so no images
have been published past a recent master commit
> - The failures had two distinct causes: run
[30054330748](https://github.com/paperclipai/paperclip/actions/runs/30054330748)
hit `ERR_PNPM_LOCKFILE_CONFIG_MISMATCH` (committed `pnpm-lock.yaml`
drifted from `patchedDependencies` in package metadata), and run
[30050197392](https://github.com/paperclipai/paperclip/actions/runs/30050197392)
hit `no space left on device` during the multi-arch buildx export
> - This pull request hardens the publish job against both failure
modes: it refreshes the lockfile (lockfile-only, guarded) before the
build, and frees runner disk space before buildx setup
> - The benefit is that image publishing keeps working through routine
lockfile drift and the growing multi-arch build footprint, so `ghcr.io`
images stay current with master
## Linked Issues or Issue Description
- Refs #8286 — same class of Docker-build lockfile mismatch failure
- Refs #8827 — pnpm 9.15.x pin / lockfile regeneration discussion
- Note: the immediate lockfile drift on master was fixed by #10132; the
refresh step here prevents the *next* drift from breaking image
publishing again
## What Changed
- Added a pnpm + Node setup and a **"Refresh lockfile for Docker build
context"** step to the image job in `.github/workflows/docker.yml`: runs
`pnpm install --lockfile-only --ignore-scripts --no-frozen-lockfile`,
exits cleanly if nothing changed, and **fails the job if anything other
than `pnpm-lock.yaml` was modified** by the refresh
- Added a **"Free runner disk"** step (before buildx setup) that prunes
the pnpm store, apt caches, preinstalled toolchains
(`/usr/share/dotnet`, Android SDK, Swift, Boost, PowerShell, GHC,
CodeQL/PyPy/Ruby toolcache), and dangling Docker state, logging `df -h`
before/after
- No changes outside the workflow file (54 added lines, nothing removed)
## Verification
- Pulled the logs of both failed master runs and matched each failure to
the step that addresses it:
[30054330748](https://github.com/paperclipai/paperclip/actions/runs/30054330748)
failed with `ERR_PNPM_LOCKFILE_CONFIG_MISMATCH`,
[30050197392](https://github.com/paperclipai/paperclip/actions/runs/30050197392)
failed with `no space left on device` during the buildx export
- Confirmed pnpm `9.15.4` in the new setup step matches the repo
`packageManager` field and the version used in the Dockerfile, so the
refreshed lockfile is generated by the same pnpm the image build
consumes
- Validated the workflow YAML parses cleanly
- The workflow triggers on master pushes / manual dispatch; the
definitive check is the first master run after merge — reviewers can
also `workflow_dispatch` it from this branch if desired
## Risks
- The lockfile refresh runs with `--ignore-scripts` and a guard that
aborts on any non-lockfile change, so it cannot silently pull unexpected
code into the image; worst case it fails the job with a clear diff
- The published image could be built from a refreshed lockfile that
differs from the committed one when drift exists — that keeps publishing
alive but can mask drift on master, which still needs the committed
lockfile fixed (as #10132 did)
- Disk cleanup removes preinstalled toolchains only on the ephemeral
runner for this job; other jobs/workflows are unaffected
- Low risk overall: additive steps in a single workflow file
## Model Used
- Claude (Anthropic) — `claude-fable-5` (Claude Code agent harness,
extended thinking, tool use). Used to diagnose the failing CI runs from
logs, author the workflow changes, and prepare this PR.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (no runtime code touched;
workflow YAML validated — see Verification)
- [ ] I have added or updated tests where applicable (n/a — CI workflow
change)
- [x] I have updated relevant documentation to reflect my changes (none
needed)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending — will confirm once
checks run)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending review pass)
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip uses GitHub Actions to keep generated lockfile changes
deterministic in CI
> - The workflow decides when to regenerate the lockfile based on
file/path changes
> - Patch changes can live under a top-level `patches/` directory, and
those changes also affect dependency resolution
> - If the workflow misses that path, CI can skip lockfile regeneration
when it should run
> - This pull request adds top-level `patches/` to the trigger so patch
updates participate in the existing lockfile regeneration flow
> - The benefit is that patch-related dependency changes continue to get
the same CI protection as the other manifest and workspace triggers
## Linked Issues or Issue Description
No public GitHub issue is linked here. The underlying problem is that
top-level `patches/` files are part of pnpm's dependency graph, but the
PR workflow's lockfile-regeneration gate only looked at package
manifests, workspace config, `.npmrc`, and `pnpmfile.*` changes. That
meant patch-only edits could skip `pnpm install --lockfile-only` and
leave downstream frozen-install jobs on a stale lockfile.
This PR keeps the existing manual lockfile edit guard in place. The
intended behavior is still: CI owns lockfile regeneration, and patch
changes are allowed to trigger that regeneration without letting
contributors commit `pnpm-lock.yaml` directly.
## What Changed
- Added top-level `patches/` to the PR workflow's dependency-resolution
trigger.
- Left the manual `pnpm-lock.yaml` edit blocker unchanged so CI still
owns lockfile regeneration.
## Verification
- `git diff --check .github/workflows/pr.yml`
- Verified the workflow path predicate matches
`patches/acpx@0.12.0.patch`, `package.json`,
`packages/shared/package.json`, `pnpm-workspace.yaml`, `.npmrc`,
`pnpmfile.cjs`, `pnpmfile.js`, and `pnpmfile.mjs`, while excluding
nested patch paths and unrelated files.
## Risks
- Low risk: this only broadens the workflow trigger set for lockfile
regeneration.
- The main behavioral change is that patch updates at the repository
root now participate in the same CI path as manifest and workspace
changes.
## Model Used
OpenAI Codex, GPT-5-based tool-using agent.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip publishes a coordinated set of packages through its
release workflows
> - Bundled packages use a pinned npm CLI so trusted publishing works
consistently
> - The canary publisher now crashes deterministically inside npm before
useful output reaches the workflow log
> - The npm debug log that contains the underlying failure disappears
with the hosted runner
> - This pull request upgrades the pinned publish CLI and preserves both
verbose HTTP activity and npm debug logs on failure
> - The benefit is that the plausible HTTP-layer fix ships immediately,
while any remaining CI-only failure becomes diagnosable
## Linked Issues or Issue Description
- **Problem:** The canary release workflow fails on the first bundled
package with `npm error Exit handler never called!` and no preceding
diagnostic output.
- **Expected behavior:** Bundled packages publish through trusted
publishing, or the workflow retains enough npm diagnostics to identify
the actual failure.
- **Reproduction:** Run the canary release workflow in GitHub Actions;
the failure reproduced on both attempts of run 29948506814.
- **Version/commit:** Current `master` after #10024 and #10030.
- **Deployment mode:** GitHub-hosted release workflow using Node.js 24
and npm trusted publishing.
- Related: #10024, #10030.
## What Changed
- Bumped the bundled publish CLI from npm 11.16.0 to npm 11.18.0.
- Added `--loglevel verbose` to bundled npm publish invocations.
- Dumped the last 300 lines of every npm debug log after failed canary
or stable publishes, with common registry credential forms redacted.
- Updated release assertions to pin npm 11.18.0 and verify verbose
logging.
## Verification
- `pnpm test:release-registry` — 66 tests passed.
- `bash -n scripts/release-lib.sh`.
- Parsed `.github/workflows/release.yml` with Python/PyYAML.
- Smoke-tested npm log redaction with representative Authorization,
`_authToken`, and token environment values.
- Smoke-tested npm debug-log redaction against Authorization,
`_authToken`, and `npm_token` examples.
- `git diff --check origin/master...HEAD`.
- The merge-triggered canary workflow remains the live
trusted-publishing verification.
## Risks
- Low code risk: changes are isolated to the release publisher and its
workflow diagnostics.
- npm 11.18.0 could expose a different registry/runtime regression;
failure-time debug log dumping makes that actionable.
- Verbose npm output increases release log volume but does not change
package contents or dist-tags; common credential forms are redacted
before debug logs are printed.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
- OpenAI Codex, GPT-5.3 Codex, tool-enabled coding agent with repository
and shell execution; context window size is not exposed in this
environment.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The pull request workflow is the main merge gate for changes to that
app.
> - The Playwright e2e lane is expensive because every spec shares one
isolated server and runs serially.
> - Splitting that lane across runners shortens the critical path, but
the public required-check contract still needs a check named exactly
`e2e`.
> - This pull request shards the real e2e work while preserving a fast
aggregate `e2e` job for branch protection.
> - The benefit is a faster PR workflow without making otherwise-good
PRs unmergeable because a legacy required check disappeared.
## Linked Issues or Issue Description
No public GitHub issue exists for this CI follow-up.
Related prior CI work:
- Refs #8360
- Refs #9168
- Refs #9516
Bug report:
### What happened?
Sharding the PR e2e lane directly at the workflow job level changes the
emitted check names to shard-specific names, while existing branch
protection expects a check named exactly `e2e`.
### Expected behavior
The PR workflow should be able to run e2e specs across multiple runners
while still emitting a stable aggregate check named `e2e`.
### Steps to reproduce
1. Open a PR against `master`.
2. Run the PR workflow with the e2e lane split only as a matrix job.
3. Observe that the shard checks complete, but a required check named
exactly `e2e` never appears.
### Paperclip version or commit
Current `master`.
### Deployment mode
GitHub Actions pull request workflow.
## What Changed
- Added `scripts/e2e-shard.mjs`, which partitions default Playwright e2e
specs by recorded per-spec duration.
- Added `scripts/e2e-shard-durations.json` with measured e2e spec
durations so the slow smoke-lab spec does not dominate one runner.
- Split the PR workflow e2e lane into two `e2e_shards` matrix jobs and
added a fast aggregate job named exactly `e2e`.
- Added `scripts/__tests__/e2e-shard.test.mjs` to lock the shard
partition, ignored-spec sync, manifest coverage, and aggregate
required-check contract.
## Verification
- `node --test scripts/__tests__/e2e-shard.test.mjs`
- `node --test scripts/__tests__/run-vitest-stable-shard.test.mjs`
- `git diff --check upstream/master..HEAD`
- Searched GitHub for duplicate or related e2e-shard / required-check
PRs and issues before opening this PR; no direct duplicate was found.
## Risks
Low risk. The main risk is that the duration manifest can drift as specs
are added or runtimes change; missing specs fall back to the median
known duration, and the focused shard test catches empty, overlapping,
or badly imbalanced partitions.
## Model Used
OpenAI GPT-5 via Codex CLI coding agent, with shell/tool execution and
repository inspection.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Paperclip publishes Docker images from this repo that self-hosters
and orchestration tooling deploy; the server refuses to start when its
database is missing any schema migration the build bundles
(`ensureMigrations`)
> - Anything that deploys these images therefore needs to know an
image's schema expectations *before* deploying it — today that requires
pulling the image or checking out the matching commit, both heavyweight
for tooling that just wants to answer "will this image boot against a
database migrated to N?"
> - Getting this wrong is expensive: an image ahead of the applied
schema crash-loops at startup and fails healthchecks after deployment
resources are already created
> - This pull request labels every published image with its bundled
migration set (last migration file and count), computed at build time
from `packages/db/src/migrations` in the same tree the image is built
from
> - The benefit is image/schema compatibility verification with two
cheap registry requests (manifest + config blob), no pull, and no drift
risk between label and image contents
## Linked Issues or Issue Description
No existing public issue; inline description per the feature request
template:
- **Subsystem affected**: Docker image publishing
(`.github/workflows/docker.yml`), `packages/db` migrations
- **Problem or motivation**: deployment tooling cannot cheaply determine
which schema migrations a published image expects; the only options are
pulling the image or checking out the matching commit. Deploying an
image whose bundled migrations exceed the applied schema makes the
server refuse to start, so this check is needed *before* resources are
created.
- **Proposed solution**: OCI labels
(`io.github.paperclipai.schema.last-migration`,
`io.github.paperclipai.schema.migration-count`), computed from the
migrations directory at build time via the existing
`docker/metadata-action` step. Keys use org-based reverse-DNS (the
GitHub org) so the label contract survives product-domain migrations.
- **Alternatives considered**: a schema manifest published beside the
image (second artifact to keep in sync — rejected); encoding schema info
in tags (tags already carry semver/sha meaning — rejected).
- **Roadmap alignment**: checked `ROADMAP.md` — no overlap with planned
core work; this is build metadata only.
## What Changed
- `.github/workflows/docker.yml`: a `Compute schema migration labels`
step (`ls` + `sort` over `packages/db/src/migrations/*.sql`) feeding two
custom labels into the existing `docker/metadata-action` step.
## Verification
- Workflow YAML validated locally.
- The label-computation commands run against the current tree produce
`last=0181_decision_training_retention_policy.sql`, `count=180`.
- After merge, verify with: fetch the image config blob for a fresh
`sha-*` tag from ghcr and confirm both `io.github.paperclipai.schema.*`
labels are present.
## Risks
- Low. Labels are metadata only; no change to image contents. If the
migrations directory ever moves, the label step fails the workflow
loudly (`ls` exits non-zero) rather than publishing wrong labels.
## Model Used
- Claude (Anthropic) — model id `claude-fable-5`, via the Claude Code
CLI harness with tool use (shell, file edits). Change authored and
verified agent-assisted, human-directed.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (YAML validation; label
commands — no app code changed)
- [x] I have added or updated tests where applicable (n/a — CI metadata
only)
- [x] I have updated relevant documentation to reflect my changes (n/a —
workflow comment documents the labels)
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Releases publish the same app and package set that operators
install, so release verification should keep full release-strength
coverage.
> - The release workflow currently verifies stable and canary releases
with one serial job that typechecks, runs all tests, and builds.
> - The PR workflow already proves the test surface can be split into
grouped general suites and serialized shards without changing coverage.
> - This pull request extracts the release verify work into a reusable
workflow and fans out the independent lanes.
> - The benefit is faster stable and canary release verification while
preserving the existing publish and preview gates.
## Linked Issues or Issue Description
No public GitHub issue exists for this CI improvement.
**Subsystem affected**
Cross-cutting (multiple of the above)
**Problem or motivation**
Release verification spends most of its wall time in a single serial
test step even though the same stable test surface is already
partitioned for PR CI. Stable dispatches and master-push canaries
therefore wait on one long runner after setup, typecheck, tests, and
build run sequentially.
**Proposed solution**
Add a reusable release verification workflow with parallel typecheck,
grouped general tests, serialized test shards, and build lanes. Have
both stable and canary release verification call it with the ref they
need to verify.
**Alternatives considered**
Keeping the serial `pnpm test:run` job preserves the old shape but keeps
stable and canary releases waiting on one long runner. Skipping
verification when a source SHA already has green CI would be faster, but
adds stale-check and lookup risk beyond this change.
**Roadmap alignment**
No overlapping item found in `ROADMAP.md`; this is release CI
maintenance.
**Additional context**
The new workflow keeps the release-strength full `pnpm -r typecheck`,
uses the existing stable test grouping/sharding entry points, and leaves
publish/preview jobs unchanged.
## What Changed
- Added `.github/workflows/release-verify.yml` as a `workflow_call`
workflow accepting a `ref` input.
- Split release verification into parallel `typecheck`, `general_tests`,
`serialized_tests`, and `build` jobs with 20-minute lane timeouts.
- Mirrored the PR workflow's stable test partition: `general-server`
shards 1-3, `general-workspaces-a`, `general-workspaces-b`, and four
serialized shards.
- Replaced `release.yml` `verify_canary` and `verify_stable` job bodies
with calls to the reusable workflow while leaving publish and preview
jobs unchanged.
- Added a Node test that guards the release workflow delegation and
split verify surface.
## Verification
- `actionlint 1.7.12 .github/workflows/release.yml
.github/workflows/release-verify.yml`
- `node ./scripts/release-package-map.mjs check`
- `node --test ./scripts/__tests__/release-verify-workflow.test.mjs
./scripts/__tests__/run-vitest-stable-shard.test.mjs`
- `git diff --check`
## Risks
- Release verification now starts more jobs per release event,
increasing total runner setup/install minutes. This matches the existing
PR CI tradeoff and should reduce release wall time substantially.
- The called workflow checks out the requested ref shallowly. That is
intentional for verify lanes; publish and preview jobs still retain
their existing full-history checkouts.
> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.
## Model Used
OpenAI Codex, GPT-5-class coding agent in local tool-use mode with shell
execution, repository editing, GitHub connector access, and medium
reasoning.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Its UI is the operator's daily surface: task lists, boards, budgets,
agent status — all built on shadcn components and Tailwind
> - Visual values (colors, spacing, type sizes, radii) were hardcoded at
~1,600 call sites: the same "small gray label" was 9/10/11px depending
on the file, charts disagreed with chips about status colors, two
toggle-switch implementations coexisted in two greens, and there was no
visual regression coverage
> - This made the UI drift-prone and made any restyle a
hundreds-of-files project, which discourages design iteration
> - This pull request extracts visual values into a single token layer
in `ui/src/index.css`, adds a Storybook visual regression suite backed
by external immutable baseline archives, and then applies a deliberate
retune reviewed change-by-change on screenshot diffs
> - The benefit is that Paperclip's look becomes a config surface:
retheming is a token edit reviewed as a snapshot diff, drift is blocked
by a token gate, and future UI PRs can prove exactly what changed
visually without committing hundreds of PNGs
## Linked Issues or Issue Description
No existing public issue covers this work (searched "design tokens",
"visual regression", "design system" across issues and PRs). Related in
spirit: Refs #8982 (theming a hardcoded panel — a one-off instance of
the same problem class this PR addresses systematically).
**Problem (feature-request form):** UI visual values are hardcoded per
call site with no source of truth and no regression coverage;
consistency depends on reviewer memory, and restyling requires mass file
edits.
**Proposed solution (this PR):** a single token layer + enforcement gate
+ externally stored visual snapshot suite, then an intentional restyle
on top of that foundation.
## What Changed
- **Token extraction (zero visual change, machine-verified during
development):** committed codemods (`scripts/codemod-*.mjs`) moved
~1,600 hardcoded color/type/spacing/radius/shadow/misc values into named
tokens in a non-inline `:root` block of `ui/src/index.css`.
- **Visual regression suite:** `pnpm test:storybook-visual` covers 255
stories × light/dark = 510 Playwright screenshots at `maxDiffPixels: 0`,
plus new primitive-coverage stories and deterministic-render fixes.
- **External visual baselines:** committed PNG snapshots were removed.
`tests/storybook-visual/baseline-manifest.json` pins an immutable
archive URL/hash/size/count, and `scripts/storybook-visual-baseline.mjs`
handles `download`, `verify`, `pack`, and trusted maintainer `upload`
flows.
- **Opt-in visual CI artifacts:** added a `Storybook Visual` workflow
that runs on manual dispatch or PRs labeled `storybook-visual`,
downloads/verifies the baseline, runs Playwright, and uploads Playwright
report/test-result artifacts for review. Normal PR runs do not mutate
baseline objects.
- **Token gate:** `pnpm check:token-gates` — zero hex literals, zero
arbitrary bracket values, zero raw font-sizes in `ui/src/components/**`
and `ui/src/pages/**`, with a documented inline allowlist for legitimate
opt-outs.
- **Theme retune (intentional, snapshot-reviewed):** new base theme
values; radius ladder derived from a single `--radius` knob; micro-type
cluster collapsed to a named ladder (`--text-nano/micro/compact` +
Tailwind `text-xs`/`text-sm`); letter-spacing collapsed to named steps.
- **One status-color vocabulary:** charts, quota/budget bar fills,
RUNNING/live chips, and liveness indicators all use the canonical
`--status-*` hues. Light-mode legibility fixes for red alert surfaces
that used dark-tuned text classes.
- **One switch:** `ToggleSwitch` restyled to the registry capsule form,
second hand-rolled implementation removed, and all call sites unified.
- **Docs:** `DESIGN.md` is the design contract; `doc/design/` holds
audit reports, decision logs, and updated guidance for external baseline
review/update workflows.
- Dead code removed (`agentStatusBadge` duplicate map), byte-identical
contrast constants consolidated, semantic renames
(`--project-seed`/`--project-none`, `--liveness-blue`).
## Verification
- `pnpm check:token-gates` — 3/3 gates CLEAN during the design-system
run
- `pnpm typecheck` && `pnpm --filter @paperclipai/ui build` — green
during the design-system run
- `node --test scripts/__tests__/storybook-visual-baseline.test.mjs` —
pass after external-baseline rework
- `pnpm exec tsc --noEmit --pretty false --module NodeNext
--moduleResolution NodeNext --target ES2022 --types
node,@playwright/test tests/storybook-visual/playwright.config.ts
tests/storybook-visual/storybook-visual.spec.ts` — pass after
external-baseline rework
- `git diff --check origin/pr/9134..HEAD` — pass after external-baseline
rework
- `find tests/storybook-visual -type f -name '*.png' -print | wc -l` —
`0`
- `node scripts/storybook-visual-baseline.mjs verify` — intentionally
fails closed until the first trusted maintainer publishes the baseline
archive and updates `baseline-manifest.json`
## Risks
- **Large but shallow:** the PR still touches many UI files due to
mechanical token extraction and retune work, but committed PNG snapshot
churn has been removed from the branch.
- **Baseline publication required before the visual suite can pass in
clean clones:** the manifest currently has placeholder archive metadata.
A trusted maintainer must publish the first immutable archive, then
update `baseline-manifest.json`.
- **Rendering platform variance:** the external baseline should be
captured in the documented Linux/Chromium environment. Future CI runs
verify against the pinned archive and fail closed on checksum/count
mismatch.
- **Visual CI is opt-in while stabilizing:** add the `storybook-visual`
label or dispatch the workflow manually to produce downloadable
Playwright report/test-result artifacts.
- **Scheduled follow-ups, deliberately out of scope:** Tailwind palette
classes map to semantic tokens in a dedicated pass; card/pill component
consolidation; ESLint ratchet. Tracked in
`doc/design/DECISION-SHEET.md`.
## Model Used
Claude Fable 5 (Anthropic, `claude-fable-5`, Mythos-class tier) with
extended thinking, running in Claude Code with tool use; mechanical
phases delegated to Claude Sonnet subagents. Follow-up external-baseline
rework assisted by OpenAI Codex (`gpt-5` coding agent with repository,
terminal, and GitHub tool use). All bulk rewrites executed via
deterministic, idempotent scripts committed in `scripts/`; intentional
visual changes were human-reviewed on screenshot contact sheets.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run targeted local verification and documented the
intentional baseline-publication failure above
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green *(pending new CI run after this
rework)*
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
*(pending review)*
- [x] I will address all Greptile and reviewer comments before
requesting merge
🤖 Generated with [Claude Code](https://claude.com/claude-code) and
OpenAI Codex
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Dotta <bippadotta@protonmail.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - CI runs a Canary Dry Run job that exercises `release.sh`, which
builds the standalone sandbox-provider packages for publish
> - That step (`scripts/build-standalone-public-packages.mjs`) built the
7 provider plugins serially — each doing `rm -rf dist && tsc` — making
it the dominant cost (~49s) inside the slowest PR check (~4.9m wall)
after the general-server lane was already sharded
> - The packages are independent (their own `node_modules` via
`--ignore-workspace`, their own `dist`), so the serial build is pure
latency with no correctness benefit
> - This pull request builds them with a bounded-concurrency pool sized
to the runner CPU count (overridable via
`STANDALONE_BUILD_CONCURRENCY`), buffering each package's output and
flushing it as one block so parallel logs stay readable, and aggregating
failures by original index
> - The benefit is a faster Canary Dry Run / PR feedback loop without
changing what gets built or published
## Linked Issues or Issue Description
No public GitHub issue exists. Inline feature/perf description:
### Problem or motivation
`build-standalone-public-packages.mjs` builds standalone provider
packages serially, making it the largest single cost inside the slowest
PR check.
### Proposed solution
Run independent per-package builds through a bounded-concurrency worker
pool sized to runner CPU count, with an env override and readable
buffered logs.
### Alternatives considered
Keep the serial build for simpler logs, but that preserves the avoidable
CI latency.
### Roadmap alignment
This is CI maintenance and does not overlap planned core roadmap work.
## What Changed
- `scripts/build-standalone-public-packages.mjs`: replaced the serial
per-package build loop with a bounded-concurrency pool (default = runner
CPU count, override via `STANDALONE_BUILD_CONCURRENCY`); per-package
stdout/stderr is buffered and flushed as a single block; failures are
aggregated by original package index so one failure neither aborts the
others mid-flight nor obscures which package broke.
- `scripts/__tests__/build-standalone-concurrency.test.mjs`: new
`node:test` unit suite covering the pool (limit respected, all items
run, ordered failure aggregation, env-override resolution).
- `.github/workflows/pr.yml`: wired the new unit test into the policy
job.
## Verification
- `node --test
./scripts/__tests__/build-standalone-concurrency.test.mjs` → 6/6 pass
- `node ./scripts/release-package-map.mjs check` → OK (29 enabled for CI
publish)
- `git diff --check origin/master..HEAD` → clean
## Risks
- Low risk. Build inputs/outputs are unchanged; only scheduling differs.
The concurrency is bounded by CPU count and overridable; output is
buffered per package so logs remain attributable. If a package fails,
all failures are still reported with their package index.
## Model Used
- Claude (Anthropic), `claude-opus-4-8`, extended thinking with tool
use.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change and contains no internal
Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Every PR runs the `PR` GitHub Actions workflow, whose `verify` gate
fans out into parallel test lanes (general tests, serialized server
route suites, build, typecheck)
> - The `General tests (server)` lane had grown into the run's critical
path: it executed all ~213 non-route server suites serially in a single
job (~7.2m of test time), more than 2x any other job
> - It runs serially because `server/vitest.config.ts` pins `maxWorkers:
1`, so server suites cannot parallelize within a single runner — the
only lever is spreading them across runners
> - This pull request shards that lane into 3 even partitions that run
on separate runners, mirroring the 4-way sharding already used for the
serialized route suites
> - The benefit is the lane drops from ~7.7m to ~2.4m/shard, cutting
overall PR wall time roughly in half (~8.5m → ~4.2m)
## Linked Issues or Issue Description
No public GitHub issue exists for this work, so the underlying issue is
described inline following the feature-request template.
### Problem or motivation
PR CI wall time had crept back up to ~8.5m. On a recent fully-green run,
the `General tests (server)` job took 7.72m — more than double any other
job and the clear critical path. Of that, 7.23m was pure test execution
(dependency install was a cached 0.27m). The job ran all server suites
that are not route/authz tests (213 files) one after another, because
the server vitest project pins `maxWorkers: 1`, making these suites
inherently serial within a single runner.
### Proposed solution
Shard the general-server lane across 3 parallel runners — the same
technique the route/authz suites already use — so the suite set is split
into even, deterministic partitions that run concurrently. Add a
regression test that proves the shards always cover the full suite set
with no gaps or overlap.
### Alternatives considered
- **Raise `maxWorkers` for the server project** to parallelize within
one runner — rejected: the server suites share process-level state
(DB/port), which is exactly why `maxWorkers: 1` is pinned.
- **Two shards instead of three** — would leave the lane at ~3.6m, still
above the next bottleneck (Canary Dry Run, ~4.1m wouldn't be the gate).
Three lands the lane comfortably below it.
- **Do nothing / accept the slow lane** — rejected: it gates every PR.
### Roadmap alignment
Developer-experience / CI tooling. Not core product roadmap work; does
not overlap with planned features in `ROADMAP.md`.
## What Changed
- `scripts/run-vitest-stable.mjs`: the `general-server` general-test
group now accepts `--shard-index` / `--shard-count`. It enumerates the
full server test set (the whole `server/src` tree, minus the route/authz
suites that already run in their own serialized shards) and splits it
deterministically by modulo. The non-sharded local invocation (`pnpm
test:run:general --group general-server`) is unchanged.
- `.github/workflows/pr.yml`: the `general_tests` matrix runs
`general-server` as 3 parallel shards (1/3, 2/3, 3/3). Workspace groups
are unchanged. The `verify` gate already aggregates the whole matrix
result, so the required check name is unaffected.
- `scripts/__tests__/run-vitest-stable-shard.test.mjs`: a `node:test`
suite asserting the 3 shards form a complete, non-overlapping partition
of the general-server set, that no route/authz suite leaks into it, and
that shard flags are rejected for the parallel workspace groups. Wired
into the `policy` job.
## Verification
- New partition test passes locally: `node --test
./scripts/__tests__/run-vitest-stable-shard.test.mjs` (3/3).
- Confirmed the 3 shards form a complete, non-overlapping partition of
all 213 files (71/71/71).
- Ran a live thin shard (3 real server suites, including one outside
`__tests__`) — 23 tests passed, confirming positional-include execution
works end to end.
- This PR's own CI is the authoritative check: all three `General tests
(server (n/3))` jobs went green on the prior run, collectively covering
every suite the old single job ran.
## Risks
- Low risk. No product code changes — only test orchestration and CI
matrix. Shard partitioning is deterministic and is now covered by an
automated test that fails if the partition ever develops a gap or
overlap. Modulo-on-sorted-filenames balances duration reasonably,
matching the approach already proven by the serialized route shards.
## Model Used
- Claude (Anthropic), `claude-opus-4-8`, extended thinking + tool use
(agentic coding via Paperclip).
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] If this change affects the UI, I have included before/after
screenshots (N/A — no UI change)
- [x] I have updated relevant documentation to reflect my changes
(inline comments explain the sharding rationale)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending this PR's run)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending review)
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
Bumps
[actions/download-artifact](https://github.com/actions/download-artifact)
from 4 to 8.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/actions/download-artifact/releases">actions/download-artifact's
releases</a>.</em></p>
<blockquote>
<h2>v8.0.0</h2>
<h2>v8 - What's new</h2>
<blockquote>
<p>[!IMPORTANT]
actions/download-artifact@v8 has been migrated to an ESM module. This
should be transparent to the caller but forks might need to make
significant changes.</p>
</blockquote>
<blockquote>
<p>[!IMPORTANT]
Hash mismatches will now error by default. Users can override this
behavior with a setting change (see below).</p>
</blockquote>
<h3>Direct downloads</h3>
<p>To support direct uploads in <code>actions/upload-artifact</code>,
the action will no longer attempt to unzip all downloaded files.
Instead, the action checks the <code>Content-Type</code> header ahead of
unzipping and skips non-zipped files. Callers wishing to download a
zipped file as-is can also set the new <code>skip-decompress</code>
parameter to <code>true</code>.</p>
<h3>Enforced checks (breaking)</h3>
<p>A previous release introduced digest checks on the download. If a
download hash didn't match the expected hash from the server, the action
would log a warning. Callers can now configure the behavior on mismatch
with the <code>digest-mismatch</code> parameter. To be secure by
default, we are now defaulting the behavior to <code>error</code> which
will fail the workflow run.</p>
<h3>ESM</h3>
<p>To support new versions of the @actions/* packages, we've upgraded
the package to ESM.</p>
<h2>What's Changed</h2>
<ul>
<li>Don't attempt to un-zip non-zipped downloads by <a
href="https://github.com/danwkennedy"><code>@danwkennedy</code></a> in
<a
href="https://redirect.github.com/actions/download-artifact/pull/460">actions/download-artifact#460</a></li>
<li>Add a setting to specify what to do on hash mismatch and default it
to <code>error</code> by <a
href="https://github.com/danwkennedy"><code>@danwkennedy</code></a> in
<a
href="https://redirect.github.com/actions/download-artifact/pull/461">actions/download-artifact#461</a></li>
</ul>
<p><strong>Full Changelog</strong>: <a
href="https://github.com/actions/download-artifact/compare/v7...v8.0.0">https://github.com/actions/download-artifact/compare/v7...v8.0.0</a></p>
<h2>v7.0.0</h2>
<h2>v7 - What's new</h2>
<blockquote>
<p>[!IMPORTANT]
actions/download-artifact@v7 now runs on Node.js 24 (<code>runs.using:
node24</code>) and requires a minimum Actions Runner version of 2.327.1.
If you are using self-hosted runners, ensure they are updated before
upgrading.</p>
</blockquote>
<h3>Node.js 24</h3>
<p>This release updates the runtime to Node.js 24. v6 had preliminary
support for Node 24, however this action was by default still running on
Node.js 20. Now this action by default will run on Node.js 24.</p>
<h2>What's Changed</h2>
<ul>
<li>Update GHES guidance to include reference to Node 20 version by <a
href="https://github.com/patrikpolyak"><code>@patrikpolyak</code></a>
in <a
href="https://redirect.github.com/actions/download-artifact/pull/440">actions/download-artifact#440</a></li>
<li>Download Artifact Node24 support by <a
href="https://github.com/salmanmkc"><code>@salmanmkc</code></a> in <a
href="https://redirect.github.com/actions/download-artifact/pull/415">actions/download-artifact#415</a></li>
<li>fix: update <code>@actions/artifact</code> to fix Node.js 24
punycode deprecation by <a
href="https://github.com/salmanmkc"><code>@salmanmkc</code></a> in <a
href="https://redirect.github.com/actions/download-artifact/pull/451">actions/download-artifact#451</a></li>
<li>prepare release v7.0.0 for Node.js 24 support by <a
href="https://github.com/salmanmkc"><code>@salmanmkc</code></a> in <a
href="https://redirect.github.com/actions/download-artifact/pull/452">actions/download-artifact#452</a></li>
</ul>
<h2>New Contributors</h2>
<ul>
<li><a
href="https://github.com/patrikpolyak"><code>@patrikpolyak</code></a>
made their first contribution in <a
href="https://redirect.github.com/actions/download-artifact/pull/440">actions/download-artifact#440</a></li>
<li><a href="https://github.com/salmanmkc"><code>@salmanmkc</code></a>
made their first contribution in <a
href="https://redirect.github.com/actions/download-artifact/pull/415">actions/download-artifact#415</a></li>
</ul>
<p><strong>Full Changelog</strong>: <a
href="https://github.com/actions/download-artifact/compare/v6.0.0...v7.0.0">https://github.com/actions/download-artifact/compare/v6.0.0...v7.0.0</a></p>
<h2>v6.0.0</h2>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="3e5f45b2cf"><code>3e5f45b</code></a>
Add regression tests for CJK characters (<a
href="https://redirect.github.com/actions/download-artifact/issues/471">#471</a>)</li>
<li><a
href="e6d03f6737"><code>e6d03f6</code></a>
Add a regression test for artifact name + content-type mismatches (<a
href="https://redirect.github.com/actions/download-artifact/issues/472">#472</a>)</li>
<li><a
href="70fc10c6e5"><code>70fc10c</code></a>
Merge pull request <a
href="https://redirect.github.com/actions/download-artifact/issues/461">#461</a>
from actions/danwkennedy/digest-mismatch-behavior</li>
<li><a
href="f258da9a50"><code>f258da9</code></a>
Add change docs</li>
<li><a
href="ccc058e5fb"><code>ccc058e</code></a>
Fix linting issues</li>
<li><a
href="bd7976ba57"><code>bd7976b</code></a>
Add a setting to specify what to do on hash mismatch and default it to
<code>error</code></li>
<li><a
href="ac21fcf45e"><code>ac21fcf</code></a>
Merge pull request <a
href="https://redirect.github.com/actions/download-artifact/issues/460">#460</a>
from actions/danwkennedy/download-no-unzip</li>
<li><a
href="15999bff51"><code>15999bf</code></a>
Add note about package bumps</li>
<li><a
href="974686ed50"><code>974686e</code></a>
Bump the version to <code>v8</code> and add release notes</li>
<li><a
href="fbe48b1d27"><code>fbe48b1</code></a>
Update test names to make it clearer what they do</li>
<li>Additional commits viewable in <a
href="https://github.com/actions/download-artifact/compare/v4...v8">compare
view</a></li>
</ul>
</details>
<br />
[](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)
Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.
[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)
---
<details>
<summary>Dependabot commands and options</summary>
<br />
You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)
</details>
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - CI/CD pipelines use GitHub Actions to automate builds and
deployments of services
> - The `agent-runtime-images` workflow uses `docker/login-action` for
authenticating with Docker registries before pushing images
> - `docker/login-action` v3 is now superseded; v4 was released with
Node 24 as the default runtime and updated internal dependencies
(`@actions/core`, AWS SDK, `@docker/actions-toolkit`)
> - This pull request bumps `docker/login-action` from v3 to v4 in the
`agent-runtime-images.yml` workflow
> - The benefit is staying on a supported runtime version and receiving
upstream security and dependency updates
## Linked Issues or Issue Description
Automated dependency bump — no upstream issue. The underlying need is
routine dependency maintenance:
- **Change type:** Dependency version bump (GitHub Actions)
- **Scope:** Single workflow file, single-line version tag change
- **Motivation:** v4 modernizes the action runtime to Node 24 and
receives updated `@actions/core` (3.x) and AWS SDK dependencies with
security patches
## What Changed
- Bumps `docker/login-action` from `v3` → `v4` in
`.github/workflows/agent-runtime-images.yml`
- v4 switches to Node 24 as the default Actions runtime (requires runner
v2.327.1+; GitHub-hosted runners satisfy this automatically)
- No changes to the action's inputs, outputs, or behavior — the API is
fully backward-compatible
## Verification
- All CI checks pass (green) on this PR
- No input/output interface changes; the login step behavior is
identical
- GitHub-hosted runners automatically meet the Node 24 runtime
requirement (runner v2.327.1+)
## Risks
Low risk. This is a one-line version bump on a widely-used, officially
maintained Docker action. The v4 release has no breaking changes to
inputs or outputs. The only new requirement (Node 24 / runner v2.327.1+)
is satisfied by GitHub-hosted runners automatically.
## Model Used
None — automated Dependabot bump. PR description updated by Claude
Sonnet 4.6 (Anthropic, 200k context, tool use enabled) to satisfy the PR
template requirement.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [ ] I will address all Greptile and reviewer comments before
requesting merge
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
> [!NOTE]
> This is **stage 3 of 3** of the staged Kubernetes contribution: stage
1 is the kubernetes sandbox-provider plugin (#5790), stage 2 is the
provider backend/hardening refresh filed separately, and this stage
ships the runtime images those sandboxes run.
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Sandboxed agent execution (Refs #248) runs each agent turn in an
isolated environment; the kubernetes sandbox provider (stage 1, #5790)
schedules those runs as hardened pods
> - A sandbox pod needs a runtime image with the harness CLI
preinstalled: installing CLIs at run start is slow, flaky, and needs
network egress the sandbox should not have
> - There is no first-party image family for this, so every deployer
would have to hand-roll Ubuntu + Node + CLI images per harness and solve
signal handling, non-root, and image chaining themselves
> - This PR ships the agent-runtime image family: a hardened base
(non-root uid 1000, tini, git, the agent shim) plus one derived image
per harness, a buildx bake file that chains them, and a publish workflow
with cosign keyless signing
> - The benefit is that any sandbox infrastructure, the kubernetes
provider or otherwise, gets ready-made, signed, security-hardened
per-harness runtime images that are verified in production across five
harnesses
## Linked Issues or Issue Description
Refs #248 (sandboxed agent execution proposal) and #5790 (the kubernetes
sandbox provider, stage 1 of this contribution, which consumes these
images as per-run runtime images via its adapter defaults).
No issue covers the image gap itself, described in-PR: sandbox providers
reference `ghcr.io/paperclipai/agent-runtime-*` images, but the
repository contains neither the Dockerfiles nor the workflow that builds
and publishes them. Without this, self-deployers cannot reproduce or
audit the images their agent runs execute in.
## What Changed
- `docker/agent-runtime/Dockerfile.base`: foundation image. Ubuntu 22.04
+ Node 22 + git + tini (PID 1, signal propagation) + non-root
`paperclip` user (uid/gid 1000) + the agent shim compiled in a Go build
stage. `WORKDIR /workspace`, entrypoint `tini -- paperclip-agent-shim`.
- One derived Dockerfile per harness: `opencode` (opencode-ai), `pi`
(@mariozechner/pi-coding-agent), `codex` (@openai/codex), `gemini`
(@google/gemini-cli, plus headless auth-mode settings), `claude`
(@anthropic-ai/claude-code, symlinked as `claude-code`). Each installs
the CLI as root, returns to uid 1000, and asserts the binary is on PATH
at build time.
- `acpx` and `hermes` Dockerfiles are included in the bake group but are
not in the default publish scope (hermes is a stub until a CLI package
exists).
- `docker/agent-runtime/buildx-bake.hcl`: builds the whole family in one
pass. Derived targets chain off the `base` target through bake
`contexts` (the literal registry in each `FROM` is overridden to
`target:base` at build time, so no intermediate push is needed).
`REGISTRY` (default `ghcr.io/paperclipai`) and `VERSION` are overridable
variables.
- `tools/agent-shim/`: a small Go shim that runs as the container
command. It reads `/run/paperclip/runtime-command.json` (`{ "command",
"args" }`), resolves the harness CLI on PATH, and `syscall.Exec`s it so
SIGTERM from the kubelet reaches the harness directly. Harness-agnostic,
with unit tests.
- `.github/workflows/agent-runtime-images.yml`: builds and pushes the
default scope (base, opencode, pi, codex, gemini, claude) for
linux/amd64 on `workflow_dispatch` (explicit version tag) or pushes to
`master` touching these paths, then signs every digest with cosign
keyless OIDC. Uses only `GITHUB_TOKEN`; no extra secrets.
- `docker/agent-runtime/README.md`: image lineup, base contents, local
build instructions, the runtime-command contract, and the security
model.
Additive only: nothing in the product loads these images. Deployments
opt in via their sandbox provider configuration (for example the
kubernetes plugin's image settings).
## Verification
- `cd tools/agent-shim && go build ./... && go test ./... && go vet
./...`: all passing.
- `docker buildx bake -f docker/agent-runtime/buildx-bake.hcl --print
base opencode pi codex gemini claude`: resolves cleanly; every tag and
build context lands on `ghcr.io/paperclipai/agent-runtime-*` and derived
targets map the base ref to `target:base`.
- Workflow YAML validated (parses, single job, no org-specific secrets).
- This exact image family (built from these Dockerfiles, bake file, and
workflow) is what runs agent execution in production on paperclip.inc,
verified end-to-end across five harnesses (opencode, pi, codex, gemini,
claude): each as a full loop from assigned issue to per-run runtime
image in a sandboxed pod to completed run.
## Risks
- Low risk: purely additive, nothing in paperclip-server or the UI
references these files. The workflow only triggers on its own paths.
- Derived images install harness CLIs `@latest` at build time; a broken
upstream CLI release would surface at image build, not at run time, and
the PATH assertion fails the build rather than shipping a broken image.
- The hermes image is an explicit stub (documented in its Dockerfile)
until a hermes CLI package exists; it is outside the default publish
scope.
- cosign signing is keyless OIDC with the workflow identity; no
long-lived signing keys are introduced.
## Model Used
Claude Opus 4.8 (claude-opus-4-8, 1M context, extended thinking, tool
use via Claude Code).
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots (no UI changes)
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The commitperclip review workflow runs a security gate as part of CI
on every PR
> - The security script's header promises it always exits 0 and stays
silent/informational, but PRs that triggered a flag were failing with a
5-minute timeout
> - Two compounding bugs: `findExistingDraftAdvisory` paginated without
an upper bound, and the workflow step did not have `continue-on-error:
true`, so any hang inside the script turned into a hard `review` check
failure that blocked merge
> - This pull request caps the advisory pagination at 20 pages and adds
`continue-on-error: true` to the workflow step, aligning runtime
behavior with the script's documented "always exit 0" contract
> - The benefit is that future PRs flagged by the security gate no
longer block merge on a 5-minute timeout, and the gate stays
silent/informational as intended
## Linked Issues or Issue Description
Fixes: #7849
## What Changed
- `.github/workflows/commitperclip-review.yml`: added
`continue-on-error: true` to the `Run security gates` step so a hang or
non-zero exit cannot fail the `review` check (matches the script's
documented "always exit 0" contract).
- `.github/scripts/check-pr-security.mjs`: capped
`findExistingDraftAdvisory` pagination at 20 pages (= 2000 advisories)
and short-circuited with a `console.warn` when the cap is hit; if no
match is found within the cap, callers will simply create a new draft
instead of hanging forever.
- `.github/scripts/tests/check-pr-security.test.mjs`: added a test
asserting the pagination cap is enforced.
## Verification
- `node .github/scripts/tests/check-pr-security.test.mjs` — 31/31 pass,
including the new cap test.
- Step-level guarantee: `continue-on-error: true` makes the `Run
security gates` step non-blocking for the job, so even an unexpected
hang/timeout in this step can no longer fail the `review` check.
## Risks
- Low risk. Pagination cap is a defensive bound; the worst case is a
duplicate draft advisory (acceptable — the workflow continues).
`continue-on-error: true` is exactly what the script header already
promised; the workflow now matches its stated contract.
## Model Used
- Claude (claude-opus-4-7), extended thinking, tool use
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [ ] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The PR pipeline runs a "Canary Dry Run" step that validates
dependency PRs against a regenerated `pnpm-lock.yaml` before merge
> - When a dependabot (or other) PR's regenerated lockfile is
byte-identical to `HEAD`, `git commit` exits non-zero with "nothing to
commit, working tree clean"
> - The surrounding `bash -e` step propagates the non-zero exit, failing
the Canary Dry Run for PRs that should be allowed through
> - This pull request guards the commit with `git diff --cached --quiet`
so the step only commits when there is a real staged change
> - The benefit is that dependabot PRs (#7571 storybook bump, #7572) and
any future PR where regen happens to match HEAD stop getting spuriously
blocked by the canary step
## Linked Issues or Issue Description
Refs #7571
Refs #7572
PR #7571 (dependabot: storybook 10.3.5 → 10.4.2) and PR #7572 both fail
on the exact same line of the `Canary Dry Run` step:
```
+ git -c user.email=ci@paperclip.local -c user.name=CI commit --no-verify -m 'ci(canary): stage regenerated lockfile'
On branch master
nothing to commit, working tree clean
Error: Process completed with exit code 1.
```
The step regenerates `pnpm-lock.yaml`, stages it, and unconditionally
commits. When the regenerated lockfile matches `HEAD` exactly (which
happens for some dependabot bumps where the lockfile resolution did not
actually change), `git commit` exits 1 and `bash -e` fails the entire
step.
## What Changed
- `.github/workflows/pr.yml`: Wrap the `git commit` inside the Canary
Dry Run step in `if ! git diff --cached --quiet; then ... fi`, so the
commit is skipped when there is no staged diff.
## Verification
- Reproduced the failure on PR #7571 and PR #7572 (identical stack trace
at the `git commit` line of the Canary Dry Run step).
- Local sanity check: `git diff --cached --quiet` returns 0 (no diff)
when the regen is a no-op and non-zero when there is a real change,
which matches the intended branching.
- Once merged, dependabot PRs that previously stalled on this step
should re-run the Canary Dry Run cleanly.
## Risks
- Low risk: the change only adds a guard before an existing `git
commit`. The `else` branch (`git checkout -- pnpm-lock.yaml` when no
artifact lockfile was used) is unchanged. No behavior change when there
*is* a regen diff.
## Model Used
- Claude (Anthropic) — `claude-opus-4-7`, extended thinking off, tool
use enabled.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have run tests locally and they pass (workflow YAML change;
verified via shell sanity check of the guard condition)
- [ ] I have added or updated tests where applicable (no test harness
for workflow YAML)
- [x] If this change affects the UI, I have included before/after
screenshots (n/a — CI workflow only)
- [x] I have updated relevant documentation to reflect my changes (n/a)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip orchestrates AI agents for zero-human companies
> - Test infrastructure across server, ui, packages/* runs on Vitest
> - Dependabot opened a narrow bump (3.2.4 → 3.2.6), but the wider
workspace is on 3.2.4 and the major-version bridge to v4 needs a
coordinated change set across configs and tests
> - Staying on 3.x indefinitely leaves us behind on Vitest 4 (perf,
pool, and config improvements) and forces repeated patch-only dependabot
churn
> - This pull request upgrades Vitest to 4.1.8 across the workspace,
updates `server/vitest.config.ts` and `scripts/run-vitest-stable.mjs`
for the new API, and adjusts two UI tests for the new assertion
semantics
> - The benefit is a single, coherent Vitest 4 upgrade that supersedes
#7570 and gets us on the supported major line
## What Changed
- Bump `vitest` from `3.2.4` to `4.1.8` across root, `server`, `ui`, and
all `packages/*` (including plugin examples and sandbox providers)
- Update `server/vitest.config.ts` for Vitest 4 config surface
- Update `scripts/run-vitest-stable.mjs` to match the new runner
behavior
- Adjust `ui/src/components/CommentThread.test.tsx` and
`ui/src/components/MarkdownEditor.test.tsx` for Vitest 4 matcher/timing
semantics
- Refresh `pnpm-lock.yaml`
## Verification
- `pnpm install` resolves cleanly with the new lockfile
- `pnpm -w -r test` (server, ui, packages) runs under Vitest 4.1.8
## Risks
- Major-version Vitest bump: behavioral changes in pools, fake timers,
and matcher strictness can surface flake. Test config and the two UI
tests were updated to match v4 semantics; broader test runs should be
watched on CI before merge.
- Supersedes dependabot PR #7570 (3.2.4 → 3.2.6); that PR should be
closed.
## Model Used
- Claude (Anthropic) — `claude-opus-4-7`, extended thinking, tool use
enabled
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [ ] I have run tests locally and they pass
- [ ] I have added or updated tests where applicable
- [ ] If this change affects the UI, I have included before/after
screenshots
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
Closes#7570
## Thinking Path
> - Paperclip orchestrates AI agents for zero-human companies
> - The CI pipeline (`.github/workflows/pr.yml`) gates every PR with a
`policy` job and a fleet of downstream test/build/release jobs
> - `policy` blocks committing `pnpm-lock.yaml` from non-dependabot,
non-`chore/refresh-lockfile` branches and (when manifests change)
already regenerates the lockfile in memory — but throws it away
> - Every downstream job runs `pnpm install --frozen-lockfile`, so
manifest-only PRs from human/agent branches deadlock: they can't commit
the lockfile and can't pass `--frozen-lockfile` either, and
`refresh-lockfile.yml` only runs `on: push: master`
> - This pull request closes the loop by uploading the regenerated
lockfile as a workflow artifact (`pr-lockfile`) from `policy` and
restoring it in each downstream job before `pnpm install
--frozen-lockfile` runs
> - The benefit is that any manifest-only PR (e.g. a routine dep bump)
goes green end-to-end without a separate refresh-lockfile round-trip,
and unrelated PRs are unaffected because the download step is
best-effort (`continue-on-error: true`)
## What Changed
- `.github/workflows/pr.yml::policy`: add `id: regen_lockfile` to the
existing manifest-detection step, emit a `regenerated` output, and
conditionally `actions/upload-artifact@v4` `pnpm-lock.yaml` as
`pr-lockfile` (retention 1 day) when a manifest changed
- `.github/workflows/pr.yml` (every downstream job —
`typecheck_release_registry`, `general_tests` [3 matrix], `build`,
`verify_serialized_server` [4 matrix], `canary_dry_run`, `e2e`): add a
`Restore regenerated PR lockfile (if policy uploaded one)` step using
`actions/download-artifact@v4` with `continue-on-error: true`, placed
before the existing `pnpm install --frozen-lockfile`
- `.github/workflows/pr.yml::canary_dry_run`: guard the pre-existing
`git checkout -- pnpm-lock.yaml` so it no longer clobbers an
artifact-restored lockfile when one is present
## Verification
- `node -e "require('js-yaml').load(...)"` parses the file; all 6
downstream jobs have the restore step, `policy` has the upload step
- This branch is itself a manifest-untouched PR, so `policy` will not
upload an artifact and every downstream job will skip the download
silently — the existing `--frozen-lockfile` path is exercised end-to-end
with no behavioral change for non-manifest PRs
- Once merged, the next manifest-touching PR (PAPA-530's #7581) will
exercise the artifact path and is expected to go green
## Risks
- **Behavior on non-manifest PRs:** the download step uses
`continue-on-error: true` so a missing artifact is silently ignored; net
behavior on PRs that don't change manifests is identical to today
- **Concurrency / cross-PR artifact bleed:** artifacts are scoped to a
single workflow run, so two concurrent PRs can't see each other's
`pr-lockfile`
- **`canary_dry_run`:** the prior unconditional `git checkout --
pnpm-lock.yaml` was a defensive no-op when `--frozen-lockfile` is
honored; the new conditional preserves that for non-manifest PRs and
additionally preserves the artifact-restored lockfile for manifest PRs
## Model Used
- Claude (Anthropic) — `claude-opus-4-7`, extended thinking, tool use
enabled
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [ ] I have run tests locally and they pass
- [ ] I have added or updated tests where applicable
- [ ] If this change affects the UI, I have included before/after
screenshots
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
## Thinking Path
> - Paperclip orchestrates AI agents for zero-human companies, including
how we ship the public `paperclip` repo itself
> - The `PR` and `commitperclip PR Review` workflows are the CI gating
layer that decides whether any pull request — human or bot — can be
merged to `master`
> - Dependabot opens dependency PRs that always carry a `pnpm-lock.yaml`
diff and an auto-generated PR body, but our `policy` job hard-fails any
non-`chore/refresh-lockfile` lockfile change, and our `commitperclip`
quality gate requires a Thinking-Path / What-Changed / Verification /
Risks / Model template Dependabot can't produce
> - Because `policy` fails first, every downstream lane (`Build`,
`Typecheck + Release Registry`, `General tests`, `Verify serialized
server`, `Canary Dry Run`, `e2e`, and the required `verify` check) skips
and `verify` fails — so we never see whether the upgrade is actually
safe
> - Socket.dev (PR Alerts + Project Report) and Snyk already run on
every dependency PR and are the supply-chain compensating control
against malicious upgrades; the missing piece is just letting our own
build/test signal run so a human can merge with confidence
> - This pull request adds a narrow Dependabot bypass to the two gates
that block on lockfile diffs and PR-template prose, while leaving every
other policy and security check active
> - The benefit is that Dependabot PRs like #7331 will now run the full
PR matrix, giving reviewers real evidence to approve or reject — without
weakening any check that targets supply-chain or build-correctness risk
## What Changed
- `.github/workflows/pr.yml` — extended the existing
`chore/refresh-lockfile` bypass on the `policy` job's "Block manual
lockfile edits" step to also skip when `github.actor ==
'dependabot[bot]'`. Every other policy step (Dockerfile deps stage
validation, `no-git-push` enforcement, release-package map check,
release bootstrap, manifest-driven `pnpm install --lockfile-only`
resolution) keeps running on Dependabot PRs.
- `.github/workflows/commitperclip-review.yml` — gated the `Run quality
gates` step and the dependent `Fail if quality gates failed` step on
`github.event.pull_request.user.login != 'dependabot[bot]'`. `Run
security gates` (`check-pr-security.mjs`) stays unconditional so
supply-chain visibility into Dependabot lockfile churn is preserved.
No changes to `.github/scripts/*.mjs` — keeping the bypass at the
workflow level avoids churning unit-tested code.
## Verification
- CI on this PR: `policy` should pass and the downstream lanes (`Build`,
`Typecheck + Release Registry`, `General tests`, `Verify serialized
server`, `Canary Dry Run`, `e2e`, `verify`) should all run normally
(this PR isn't from Dependabot, so the bypass condition is false —
proves we didn't accidentally widen the exemption).
- After merge, ask Dependabot to rebase #7331 (`@dependabot rebase`) and
confirm:
- `PR / policy` → `success` (lockfile step now `skipped`, other policy
steps `success`)
- `PR / Build`, `PR / Typecheck + Release Registry`, `PR / General tests
(server|workspaces-a|workspaces-b)`, `PR / Verify serialized server
(1/4..4/4)`, `PR / Canary Dry Run`, `PR / e2e` → all execute (none
`skipped`)
- `PR / verify` → `success` once the matrix passes
- `commitperclip PR Review / review` → `success` (quality-gates steps
`skipped` for Dependabot; security gates ran)
- Socket and Snyk checks unchanged
- Local sanity-check: `git diff origin/master..HEAD` shows only the two
workflow files, 7 added / 2 removed lines.
## Risks
- **Auto-merging a poisoned dep.** Mitigated by Socket.dev + Snyk +
human merge approval. This change only affects CI gating, not who clicks
"Merge".
- **Spoofing `github.actor` as `dependabot[bot]`.** GitHub sets
`github.actor` from the push actor; spoofing requires a compromised
Dependabot install token, which is the same threat model that already
lets an attacker push anything to a Dependabot-controlled branch — not a
new risk surface.
- **Policy "Validate dependency resolution when manifests change" step
running `pnpm install --lockfile-only --no-frozen-lockfile` on a
Dependabot lockfile.** That step intentionally uses `--lockfile-only`,
so it only verifies the manifest resolves and does not push or commit
the result. Existing behavior is unchanged.
- Low overall: the diff is two workflow-level `if:` conditions in steps
that already had bypasses.
## Model Used
- Provider: Anthropic Claude (via Claude Code in the Paperclip executor)
- Model ID: claude-opus-4-7
- Context window: 200K
- Reasoning mode: standard tool-use; no extended thinking required for
this change
- Capabilities used: file edit, bash, GraphQL/REST API calls
- Plan was drafted, approved by board, and split into child issues
before implementation; see
[PAPA-490](https://paperclip.ing/PAPA/issues/PAPA-490) for the planning
thread.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass (this change is
workflow-only — no code under test; lint via `yamllint` clean)
- [x] I have added or updated tests where applicable (workflow gating;
no script changes, no unit-testable surface)
- [x] If this change affects the UI, I have included before/after
screenshots (no UI changes)
- [x] I have updated relevant documentation to reflect my changes (no
docs reference these gates)
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
Fixes#6470
## Thinking Path
> - Paperclip is an open-source AI agent platform receiving a high
volume of community PRs — currently 2,398 open
> - The contributor experience is broken: PRs sit for months with no
feedback, contributors don't know why they're stuck, and maintainers
spend review time on PRs that are missing basics
> - Common problems: no linked issue, no test coverage, incomplete PR
template, manually-edited lockfile — all catchable before human review
> - At the same time, accepting untrusted PRs from unknown contributors
is a real attack surface: malicious packages, secret injection,
tampering with CI scripts, and code touching the sensitive paths from
the April security advisories
> - This PR adds automated gates that run on every PR: quality failures
get a clear comment telling contributors exactly what to fix, security
concerns are silently flagged as draft advisories and block merge via a
pending check run
> - The benefit is a dramatically faster feedback loop for good-faith
contributors and a meaningful security layer for the maintainers
reviewing them
## What Changed
- **`.github/workflows/commitperclip-review.yml`** — new workflow using
`pull_request_target` (runs in base branch context, has secrets, never
executes PR code). Runs quality gates + security gates on every PR
open/update.
- **`.github/dependabot.yml`** — weekly automated dependency
vulnerability PRs for npm and GitHub Actions.
- **`.github/scripts/get-bot-token.mjs`** — generates a short-lived
commitperclip installation token from `COMMITPERCLIP_KEY` secret.
- **`.github/scripts/run-quality-gates.mjs`** — orchestrates 5 quality
gates, posts/updates a single consolidated comment on the PR.
- **`.github/scripts/check-pr-template.mjs`** — validates all 5 required
template sections, Thinking Path depth (≥3 sentences), Model Used not
placeholder.
- **`.github/scripts/check-pr-linked-issue.mjs`** — requires `Fixes
#NNN` or issue URL in PR body.
- **`.github/scripts/check-pr-test-coverage.mjs`** — requires at least
one test file in the diff.
- **`.github/scripts/check-pr-lockfile.mjs`** — blocks manual
`pnpm-lock.yaml` edits (only the refresh bot may change it).
- **`.github/scripts/check-pr-dependencies.mjs`** — informational
comment when new npm packages are added.
- **`.github/scripts/check-pr-security.mjs`** — 6 silent security
checks: secret patterns, CI workflow tampering, build script changes,
supply chain (new packages in lockfile), suspicious test patterns
(outbound network/shell exec/env var reads), and changes to the 9
sensitive path prefixes from the April advisories. When any fire:
creates a draft security advisory + sets `security-review` check to
`in_progress` (blocks merge). When clean: sets `security-review` to
`success`.
- **`actions/dependency-review-action@v4`** — per-PR dependency
vulnerability check (fails if new dep has known CVE).
- **44 unit tests** across all gate modules (`node:test`, no external
deps).
## Verification
Run all unit tests locally:
```bash
node --test .github/scripts/tests/*.test.mjs
```
Expected: 44 pass, 0 fail.
End-to-end: open a PR missing the template, linked issue, and test files
→ commitperclip posts a consolidated comment listing all failures. Open
a PR with all gates satisfied → `✅ All checks passing` comment posted,
all check runs green.
## Risks
**`pull_request_target` security model:** This workflow runs in base
branch context and has access to secrets. It explicitly checks out `ref:
master` (never PR code) and reads the PR diff via GitHub API only — no
PR code is ever executed. This is the correct pattern for running
secret-bearing checks on fork PRs; deviating from it (e.g. checking out
the PR branch) would be a security vulnerability.
**False positives on security gates:** The sensitive-path gate flags any
PR touching the 9 path prefixes from the April advisories. Legitimate
fixes to those paths will trigger draft advisories. This is intentional
— those paths warrant a human look regardless. The `security-review`
check can be manually resolved by a maintainer once reviewed.
**commitperclip not yet installed:** Until the app is installed on this
repo and the `COMMITPERCLIP_KEY` secret is added, the workflow will fail
on the token generation step. The quality gate comment won't post, but
Dependency Review will still run independently.
## Model Used
Claude Sonnet 4.5, 200k context window, extended thinking enabled, tool
use: read/edit files, bash execution, GitHub API calls
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass (44/44)
- [x] I have added or updated tests where applicable (44 unit tests
across all gate modules)
- [ ] If this change affects the UI, I have included before/after
screenshots (N/A — CI only)
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
---
## One-time setup needed from you, Dotta
1. **Install commitperclip app** on this repo:
https://github.com/apps/commitperclip/installations/new
2. **Add `COMMITPERCLIP_KEY`** as a repository secret (Actions →
Secrets) — ask @brandonburr for the key
3. **Add `security_advisories: write` and `checks: write`** to the
commitperclip app permissions (commit-capital org → Settings → Apps →
commitperclip → Permissions)
4. **Install Socket.dev** from GitHub Marketplace for supply chain
scanning
5. **Branch protection** (optional but recommended): require
`commitperclip-review` and `security-review` checks to pass before merge
## Dashboard integration note
The `commitperclip-review` check run result maps cleanly to your PR
triage dashboard. A single filter on your Worker:
```javascript
const gatesCheck = checkRuns.find(r => r.name === 'commitperclip-review');
if (gatesCheck?.conclusion === 'failure') return null; // filter from queue
```
For security flags: `GET
/repos/paperclipai/paperclip/security-advisories?state=draft` — advisory
titles include `PR #NNN` for cross-referencing. PRs with a matching
draft advisory have `security-review` in `in_progress` state (grey
spinner, can't merge via branch protection).
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Devin Foley <devin@devinfoley.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip orchestrates AI agents across isolated execution
workspaces; the local cwd is the only persistence boundary between runs.
> - Workspace lifecycle (worktree_prepare → execute →
workspace_finalize) and the wake/accept flow are what guarantee that
dependent issues see a consistent worktree.
> - PAPA-380 / PAPA-431 / PAPA-432 / PAPA-440 surfaced three holes in
that contract: silent env reuse across assignees, dependent wakes firing
before finalize, and `issue.interaction.accept` advancing before
finalize landed.
> - PAPA-441 / PAPA-442 then needed to document the "no remote git"
contract and prevent future adapter/runtime code from quietly
reintroducing `git push` as a backdoor sync.
> - This pull request lands those server fixes, the static
`check-no-git-push` enforcement, the AUTHORING.md cross-link, and the
Cody-review follow-ups on the PAPA-430 thread.
> - The benefit is that finalize is a real barrier — board accepts,
dependent wakes, and operator-set env all respect it — and adapter code
can't bypass it via raw `git push`.
## What Changed
- **server (PAPA-380, PAPA-431):** `execution-workspace-policy` refuses
silent env reuse when the assignee's resolved env disagrees with the
workspace it would inherit. The inheritance protection is now scoped to
the actual inheritance signal — explicit issue-level `environmentId` is
honored even when the agent's default env is `null`.
- **server (PAPA-432):** `heartbeat.ts` gates dependent wakes on
`listUnfinalizedExecutionWorkspaceIds`, and writes a
`workspace_finalize` row on the succeeded path. Write failures now
surface instead of being swallowed so dependents aren't silently
stranded behind a missing row.
- **server (PAPA-440):** `issue-thread-interactions.acceptInteraction`
adds a workspace_finalize precondition for `request_confirmation` (not
`suggest_tasks`). Accept returns 409 if finalize hasn't succeeded for
the latest workspace operation.
- **ci (PAPA-442):** new `scripts/check-no-git-push.mjs` static check
scans `packages/adapters/`, `packages/adapter-utils/`, `server/src/`,
and `cli/src/` for any `git push` invocation (string or args-array).
Wired into the `policy` PR job and `test:release-registry`. Operators
can opt in per-call with `// paperclip:allow-git-push: <reason>`.
Release scripts are out of scope by design.
- **docs (PAPA-441):** `AUTHORING.md` documents the no-remote-git
contract and cross-links the static check so adapter authors learn the
rule and the enforcement together.
- **review follow-up (PAPA-430, Cody):** three fixes — env resolver bug,
accept-gate scope (request_confirmation only), and finalize record write
on the succeeded path.
## Verification
- `pnpm exec vitest run
server/src/__tests__/execution-workspace-policy.test.ts
server/src/__tests__/issue-thread-interactions-service.test.ts` → 33/33
pass
- `node scripts/check-no-git-push.test.mjs` → check covers string form,
args-array form, comment exclusions, and per-line allow-comment.
- Manual: server compiles; the policy job runs the check in <1s before
heavier jobs.
## Risks
- **Behavioral shift in accept:** boards accepting
`request_confirmation` while finalize is in-flight now get 409s. This is
intentional — they can retry — but it changes timing on a hot path.
`suggest_tasks` is unaffected.
- **Workspace policy:** the env-reuse refusal is a new error path.
Issues that previously silently reused an env from a different-assignee
workspace will now fail-loud; the resolver still honors explicit
issue-level `executionWorkspaceSettings.environmentId`.
- **CI rule:** any future legitimate `git push` in scoped dirs must be
marked with the allow-comment, which is the intended ergonomic.
## Model Used
- Claude Opus 4.7 (`claude-opus-4-7`, extended thinking), via Claude
Code in the Paperclip executor adapter.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] If this change affects the UI, I have included before/after
screenshots (N/A — server/CI/docs only)
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
Closes related issues: PAPA-430, PAPA-380, PAPA-431, PAPA-432, PAPA-440,
PAPA-441, PAPA-442
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip relies on CI browser suites to protect control-plane
workflows, so a stalled browser bootstrap is a release blocker even when
app code is unchanged.
> - The failing signal on [PAPA-457](/PAP/issues/PAPA-457) was specific
to the PR e2e lane timing out before tests started, which pointed at
environment setup rather than assertions.
> - The first shell-only Chromium attempt reduced download size, but the
GitHub Actions log showed Playwright still hanging inside its install
step after the headless shell download finished.
> - That means the real problem is the Playwright browser-install path
itself on the hosted Ubuntu runner, not just the size of the downloaded
artifact.
> - GitHub's Ubuntu runners already ship Google Chrome, and Playwright
can target that binary through the `chrome` channel without downloading
its own Chromium bundle.
> - The safer workflow fix is therefore to remove the Playwright install
step from the affected headless jobs and make the Playwright configs
optionally use runner Chrome only when CI opts into it.
> - This keeps local defaults unchanged, removes the failing
browser-download dependency from CI, and preserves headless coverage for
PR, standalone e2e, and release-smoke workflows.
## What Changed
- Updated `.github/workflows/pr.yml`, `.github/workflows/e2e.yml`, and
`.github/workflows/release-smoke.yml` to stop downloading Playwright
browsers and instead verify the runner's preinstalled `google-chrome`.
- Passed `PAPERCLIP_PLAYWRIGHT_CHANNEL=chrome` into the headless PR,
standalone e2e, and release-smoke test steps so those jobs explicitly
use runner Chrome.
- Updated `tests/e2e/playwright.config.ts` and
`tests/release-smoke/playwright.config.ts` to honor
`PAPERCLIP_PLAYWRIGHT_CHANNEL` while keeping the default
local/browser-bundle behavior unchanged when the env var is absent.
## Verification
- Investigated the failed PR run log and confirmed the prior `Install
Playwright` step stalled after `chromium-headless-shell` reached 100%
download.
- `PLAYWRIGHT_BROWSERS_PATH="$(mktemp -d)"
PAPERCLIP_PLAYWRIGHT_CHANNEL=chrome PAPERCLIP_E2E_SKIP_LLM=true pnpm run
test:e2e`
Result: `7 passed (21.1s)` with an empty temporary Playwright browser
cache, proving the e2e suite runs without any Playwright browser
download when the `chrome` channel is selected.
- `git diff --check`
## Risks
- This assumes GitHub's Ubuntu runner continues to ship `google-chrome`;
if that image contract changes, these workflows would need a dedicated
Chrome install step.
- The `chrome` channel can differ slightly from Playwright-managed
Chromium, so the config gate is intentionally env-scoped to CI workflows
that need the hosted-runner path.
## Model Used
- OpenAI Codex, GPT-5-based coding agent running through Paperclip's
`codex_local` adapter with tool use, shell execution, and repository
editing enabled. The exact internal snapshot/version string is not
exposed in-session.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass
- [ ] I have added or updated tests where applicable
- [ ] If this change affects the UI, I have included before/after
screenshots
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip orchestrates AI agents through a control-plane repo that
relies on GitHub Actions as part of its release and verification safety
net.
> - The PR workflow in `.github/workflows/pr.yml` is the core CI path
protecting pull requests before merge.
> - Baseline measurement work in [PAPA-335](/PAPA/issues/PAPA-335)
showed the old single `verify` job was the critical-path bottleneck,
with general tests and build serialized together.
> - Follow-up implementation in [PAPA-338](/PAPA/issues/PAPA-338) and
[PAPA-339](/PAPA/issues/PAPA-339) split that work into parallel lanes
and removed redundant clean-runner prebuild work.
> - [PAPA-340](/PAPA/issues/PAPA-340) now needs real post-change PR
workflow evidence, not local inference, to compare against the May 15,
2026 baseline and decide whether phase-2 work is still justified.
> - This pull request publishes the already-implemented CI speedup
branch so GitHub can run the actual `PR` workflow against it.
> - The benefit is that CI timing decisions are based on measured runs
from the exact workflow shape we intend to ship.
## What Changed
- Split the PR workflow so `policy` fans out into separate `Typecheck +
Release Registry`, grouped `General tests`, and `Build` jobs.
- Kept the serialized server matrix, canary dry run, and e2e jobs intact
while removing the old monolithic `verify` bottleneck.
- Reworked grouped general-test execution in
`scripts/run-vitest-stable.mjs` so the workflow can run balanced
non-serialized lanes.
- Replaced redundant clean-runner prebuild gates with the idempotent
`ensure-build-deps` path used by the relevant CI entrypoints.
## Verification
- `ruby -e "require 'yaml'; YAML.load_file('.github/workflows/pr.yml');
puts 'yaml-ok'"`
- `node scripts/run-vitest-stable.mjs --mode general --dry-run`
- `node scripts/run-vitest-stable.mjs --mode general --group
general-server --dry-run`
- `node scripts/run-vitest-stable.mjs --mode general --group
general-workspaces-a --dry-run`
- `node scripts/run-vitest-stable.mjs --mode general --group
general-workspaces-b --dry-run`
- `pnpm test:run:general -- --group general-workspaces-b`
- `pnpm test:run:general -- --group general-workspaces-a`
- `pnpm test:run:general -- --group general-server`
- `pnpm run typecheck:build-gaps`
- `pnpm --filter @paperclipai/plugin-hello-world-example typecheck`
## Risks
- Required-check and branch-protection settings may still reference the
old single `verify` job name.
- Parallel CI lanes can expose hidden ordering assumptions or
clean-runner bootstrap gaps that local grouped dry-runs did not surface.
- Because the branch is behind current `master`, merge conflicts or
unrelated upstream drift could affect the measured runtime until the
branch is rebased.
> Checked `ROADMAP.md`; this work is CI throughput maintenance for the
existing PR verification path, not duplicate feature work.
## Model Used
- OpenAI Codex via Paperclip `codex_local`, GPT-5-class coding agent
with repository read/write, shell execution, and GitHub CLI/tool use.
The runtime does not expose a more specific backend model ID in-session.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [ ] If this change affects the UI, I have included before/after
screenshots
- [ ] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip orchestrates AI agents for autonomous companies, so
developer throughput on the control plane repo directly affects how fast
the product can evolve.
> - The PR workflow is part of that throughput surface because every
change waits on it before review and merge.
> - This branch started from measured evidence that the PR critical path
was dominated by work that was either serialized unnecessarily or placed
on the wrong part of the graph.
> - The biggest concrete problems were: the canary dry run living inside
`verify`, the server isolated suites running one-by-one in a single
lane, and duplicate CI work that the PR path was paying for without
increasing coverage proportionally.
> - This pull request restructures the PR workflow so those costs are
reduced without removing the important coverage that was already
protecting release and test quality.
> - Follow-up fixes on the branch hardened the new entrypoints so they
work on clean GitHub runners and so the reduced PR typecheck path stays
self-maintaining as workspace packages evolve.
> - The benefit is materially faster PR wall-clock time while keeping
canary packaging checks, serialized-suite isolation, plugin SDK
consumers, and explicit TypeScript coverage where builds do not already
provide it.
## What Changed
- Moved the PR canary dry run into its own `Canary Dry Run` job so it
still runs on PRs but no longer extends the `verify` critical path.
- Split the custom Vitest runner into `general`, `serialized`, and `all`
modes, and added shard support for the isolated server suites.
- Added `test:run:general` and `test:run:serialized` scripts, then
rewired PR CI to fan the serialized server suites out across a 4-way
matrix.
- Added the required `@paperclipai/plugin-sdk` build preflight before
the new reduced-scope typecheck and test entrypoints so they succeed on
clean CI runners.
- Replaced the hardcoded PR build-gap list with
`scripts/run-typecheck-build-gaps.mjs`, which discovers workspace
packages whose `build` scripts skip TypeScript and runs only their
explicit `typecheck` scripts.
- Removed the redundant `pnpm build` from the PR `e2e` job because the
Playwright onboarding path boots Paperclip from source.
## Verification
- `ruby -e "require 'yaml'; YAML.load_file('.github/workflows/pr.yml');
puts 'workflow ok'"`
- `node scripts/run-vitest-stable.mjs --mode general --dry-run`
- `node scripts/run-vitest-stable.mjs --mode serialized --shard-index 0
--shard-count 4 --dry-run`
- `pnpm run typecheck:build-gaps`
- `pnpm test:run:general`
- `pnpm test:run:serialized -- --shard-index 0 --shard-count 4`
- `pnpm build`
- `pnpm paperclipai onboard --yes --run`
- `curl http://127.0.0.1:3299/api/health`
## Risks
- Branch protection or required-check configuration may need to be
updated for the new standalone `Canary Dry Run` job and the
serialized-suite matrix job names.
- `scripts/run-typecheck-build-gaps.mjs` assumes packages that need
explicit PR-time typechecking are the ones whose `build` scripts omit
`tsc`; if build conventions change, that heuristic needs to stay
aligned.
- Serialized test sharding preserves per-suite isolation, but the first
few CI runs should still be watched for shard-balance or naming
assumptions in downstream tooling.
## Model Used
- OpenAI GPT-5.4 via the Codex local adapter, using high reasoning
effort with shell, git, and file-edit tool use in a local worktree.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is a control plane for autonomous agent companies, so its
release automation is part of the core operator trust boundary.
> - The affected subsystem is npm/GitHub Actions release publishing for
the public monorepo packages.
> - The concrete failure was that a newly added package reached
`master`, the canary workflow attempted its first publish, and npm
trusted publishing was not yet bootstrapped for that package.
> - That means the problem is not just one broken run; it is a missing
pre-merge guard that lets release-ineligible packages land and only fail
once `publish_canary` runs.
> - This pull request makes release enrollment explicit, validates that
enrollment in CI, and adds a PR-time bootstrap check against npm for
changed release-enabled package manifests.
> - The result is that we keep trusted publishing, avoid teaching CI to
`npm adduser`, and move this class of failure from post-merge canary
time to pre-merge review time.
## What Changed
- Added `scripts/release-package-manifest.json` so release-managed
public packages are explicitly enrolled instead of being inferred from
every non-private workspace package.
- Hardened `scripts/release-package-map.mjs` to validate the manifest
before release workflows rewrite versions or assemble publish payloads.
- Added `scripts/check-release-package-bootstrap.mjs` and wired it into
`.github/workflows/pr.yml` so PRs that change a release-enabled package
manifest fail if that package does not already exist on npm.
- Added release-package manifest coverage tests to
`scripts/release-package-map.test.mjs` and included them in `pnpm run
test:release-registry`.
- Wired manifest validation into `.github/workflows/release.yml` and
documented the first-publish bootstrap policy in `doc/PUBLISHING.md` and
`doc/RELEASE-AUTOMATION-SETUP.md`.
## Verification
- `pnpm run test:release-registry`
- `./scripts/release.sh canary --skip-verify --dry-run`
- Confirmed the committed diff contains no obvious PII/secrets via
targeted pattern scan before pushing.
## Risks
- Low risk overall: this is CI/release-policy code, not product runtime
logic.
- The new PR bootstrap check depends on npm metadata availability, so a
transient npm outage could block a PR that changes a release-enabled
package manifest.
- The manifest introduces a new source of truth that must stay aligned
with public package additions, but that is intentional and now enforced.
## Model Used
- OpenAI Codex via the `codex_local` Paperclip adapter; GPT-5-based
coding agent with tool use, terminal execution, git, and GitHub CLI.
Exact served model ID/context window are not exposed by the local
runtime.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
Dotta
Devin Foley
Nicky Leach
Zannis Kalampoukis
Tonio
Andrew Aymeloglu
dependabot[bot]
scotttong
Jannes Stubbemann
brandonburr
Dotta