## Thinking Path
> - Paperclip is the open-source app people use to manage AI agents for
work
> - The `ui` package contains React components that drive the agent
scheduling and workspace configuration UX
> - Dependabot PR #9895 bumped `@radix-ui/react-*` from 1.6.0 → 1.6.4,
which changed the internal effect-scheduling order inside Dialog and
Select primitives
> - Two `workspaces-a` CI tests started failing:
`RoutineRunVariablesDialog` and `editable-sections / TriggersSection`
> - Root cause for both: radix 1.6.4's changed scheduling pushes a
cascaded state update one render-tick later, and each test asserted
before that tick landed
> - This PR fixes the two affected surfaces at the source (one
production fix, one test fix) so the radix bump can land cleanly
> - The benefit is unblocking PR #9895 without compromising test
fidelity or production correctness
## Linked Issues or Issue Description
Refs #9895 — this PR fixes the two `workspaces-a` test failures that
blocked the radix-ui 1.6.0 → 1.6.4 Dependabot bump.
**Root cause:** radix-ui 1.6.4 changed the internal effect-scheduling
order inside its Dialog and Select primitives, pushing certain cascaded
state updates one render-tick later than before. Two tests each asserted
against the intermediate state, before the deferred tick landed.
**Affected tests (both now pass at radix 1.6.0 AND 1.6.4):**
1. `editable-sections / TriggersSection` — `ScheduleEditor`
disabled-button assertion
2. `RoutineRunVariablesDialog` — workspace branch propagation assertion
Production behavior is unchanged and correct in both cases (verified via
tracing).
## What Changed
- **`ui/src/components/ScheduleEditor.tsx`** — `onValidityChange` is now
called synchronously inside the custom-cron `onChange` handler, not only
via the passive `useEffect` below. Previously there was a one-render
window where an invalid draft still read as valid to the parent (button
enabled); this closes that window. The effect call is preserved as a
safety net for other entry paths; only the `onChange` path is new.
- **`ui/src/components/RoutineRunVariablesDialog.test.tsx`** — the
post-mount settle loop now waits for the branch value to actually appear
in an `<input>` (`value === "pap-1634-routine-branch"`) rather than
exiting as soon as the workspace card mounts. The card reports its
branch name through an effect callback that triggers a follow-up render;
the old loop exited one tick too early. Iteration cap raised from 10 →
20 to give the extra tick room.
This PR intentionally does **not** bump radix-ui — that stays in #9895.
## Verification
```sh
# TypeScript — clean at radix 1.6.0 (master):
tsc -p ui/tsconfig.json
# Targeted vitest:
npx vitest run ui/src/components/RoutineRunVariablesDialog.test.tsx
npx vitest run ui/src/components/editable-sections
npx vitest run ui/src/components/ScheduleEditor
# Full ui suite — green with radix 1.6.4 installed locally (371 files / 3035 tests):
npx vitest run --project ui
# check-forbidden-tokens — clean
```
All of the above pass at **both** radix 1.6.0 (current master) and
1.6.4.
## Risks
Low risk. The production change (`ScheduleEditor.tsx`) adds a
synchronous call to an already-injected `onValidityChange` prop — same
value, earlier in the same event cycle. No new state, no new effects, no
API changes. The test change tightens an assertion (waits longer, checks
a more specific condition) rather than relaxing one.
## Model Used
Claude Sonnet (Anthropic) — Paperclip agent workflow; model family
claude-sonnet-4-x with tool use and extended reasoning 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: Harold Kim <harold@paperclip.ing>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Agents and board users can attach files to issues so context and
deliverables stay with the task
> - Some clients upload Microsoft Office files with generic binary MIME
types such as `application/octet-stream`
> - Current `master` now accepts arbitrary issue attachment MIME types,
so the upload should keep working for unknown binary files
> - Office files still benefit from being stored with a specific Office
MIME type when the filename makes that inference safe
> - Shared attachment allow-list defaults should also include common
Office MIME types for routes that still use that allow-list
> - This pull request keeps the current arbitrary-MIME issue upload
behavior and only narrows generic binary uploads to Office MIME types
for known Office filename extensions
## Linked Issues or Issue Description
Fixes#8243
Duplicate search performed before implementation:
- No matching open or closed PR found for `8243`, `Office document`,
`attachment MIME`, or `openxmlformats`.
## What Changed
- Added common Office MIME types to the default shared attachment
allow-list.
- Added upload content-type normalization that maps generic binary
uploads to a specific Office MIME type only for known Office filename
extensions.
- Added an optional helper-level allow-list gate so callers that still
validate against an effective allow-list can keep generic binary uploads
generic when the inferred Office MIME type is not allowed.
- Reused the shared generic attachment content-type list for response
handling.
- Preserved current `master` behavior for issue uploads that use unknown
or arbitrary MIME types.
- Added regression coverage for default Office allow-list matching,
filename inference, optional allow-list fallback, official Office MIME
uploads, inferred generic Office uploads, and preservation of unknown
generic binary uploads.
## Verification
- `env CI=true corepack pnpm install --frozen-lockfile --force`
- `env CI=true corepack pnpm --filter @paperclipai/server exec vitest
run src/__tests__/attachment-types.test.ts
src/__tests__/issue-attachment-routes.test.ts`
- `env CI=true corepack pnpm --filter @paperclipai/plugin-sdk
ensure-build-deps`
- `env CI=true corepack pnpm --filter @paperclipai/server exec tsc
--noEmit`
- `git diff --check origin/master...HEAD`
GitHub CI, security checks, and Greptile pass on rebased head
`acc364cfbe3440a59db6570bb907818046649eb4`.
## Risks
Low risk. The issue attachment route continues to accept arbitrary MIME
types as current `master` does; this change only stores a more specific
Office MIME type for generic binary uploads when the filename has a
known Office extension. Unknown generic binary uploads remain generic.
For callers that use an allow-list before storing uploads,
`normalizeUploadAttachmentContentType` supports an optional gate so
inference can be limited to MIME types that are already allowed.
No docs change included because this is a default upload compatibility
fix covered by server tests.
> 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`.
This is a narrow bug fix, not roadmap-level core feature work.
`ROADMAP.md` was checked.
## Model Used
OpenAI Codex using GPT-5, tool-enabled coding agent. Context window
details are 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: Sami Rusani <sr@samirusani>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Issues can carry an embedded multi-stage `executionPolicy` (e.g. QA
→ CodeReviewer → CodePusher) driven by
`applyIssueExecutionStageTransition` in
`server/src/services/issue-execution-policy.ts`
> - On approval, the next stage was picked with `nextPendingStage()`,
which scans the **whole** stage list from index 0 for the first id not
in `completedStageIds`
> - Stage ids are regenerated whenever the embedded policy is re-sent or
edited mid-flow (a supported operation — the existing "reassigns the
active stage when the current participant is removed" test depends on
it), so earlier `completedStageIds` can stop matching the current
policy; a final-stage approve then "finds" stage 1 pending again and
rebuilds a first-stage review (#7893) — an endless re-review loop that
can recycle indefinitely against a moving main tip
> - This pull request makes approvals advance with a forward-only scan
(only stages *after* the one being approved), so approving the last
stage always terminates the policy, and adds a guard so an
already-completed execution state is terminal for `status=done`
> - The benefit is final-stage approvals close the issue as the policy
intends, with no behavior change for non-final advancement or
reject/changes_requested verdicts
## Linked Issues or Issue Description
Fixes#7893
## What Changed
- `server/src/services/issue-execution-policy.ts`:
- New `nextPendingStageAfter(policy, completedStage, state)` helper —
forward-only scan from the approved stage's index; the approval path
uses it instead of `nextPendingStage()`. Approving the final stage
therefore always yields `nextStage === null` → completed state → the
caller's `done` flows through.
- New guard: `requestedStatus === "done"` with an already-`completed`
execution state returns without restarting the chain at stage 1 (closes
the same loop when a stale completed state lingers).
- Reject/`changes_requested` verdicts and intact-state forward
advancement are untouched.
- `server/src/__tests__/issue-execution-policy.test.ts`: 4 regression
tests, including one that reproduces the exact rewind (regenerated stage
ids + final-stage approve → previously reassigned QA at
`currentStageIndex 0`; now terminal completed) and an explicit
final-stage rejection test pinning the unchanged path.
## Verification
- `npx vitest run server/src/__tests__/issue-execution-policy.test.ts` →
54 passed (50 pre-existing + 4 new).
- `pnpm --filter @paperclipai/server typecheck` → clean.
- The rewind was confirmed empirically against unmodified code first (a
test asserting the buggy output passed pre-fix and flips post-fix), plus
brute-forced realistic operation sequences (checkout dances, status
round-trips, interim comments per the agent flow documented around
#4889) to verify intact-state flows are unaffected.
- Related suites (`issue-execution-policy-routes`,
`issue-comment-reopen-routes`, `issues-service`,
`issue-thread-interaction-routes`,
`issue-agent-mutation-ownership-routes`) also pass locally.
## Risks
- Behavior deliberately preserved: non-final approvals (forward scan is
identical when state is intact), rejections at any stage,
reopen-from-done (state cleared on reopen, fresh chain still starts at
stage 1), and explicit `in_review` restarts.
- The policy schema has no terminal-state field, so per the issue's Ask
the policy simply terminates and the requested `done` status flows
through.
## Model Used
- Claude Fable 5 (`claude-fable-5`, Anthropic) via Claude Code, agentic
mode with tool use (subagent implementation + independent adversarial
review subagent), extended thinking 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 (none found for #7893)
- [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 (N/A — server-only change)
- [x] I have updated relevant documentation to reflect my changes (N/A —
internal stage-advance semantics)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (will confirm once CI runs on
this PR)
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(pending first review)
- [x] I will address all Greptile and reviewer comments before
requesting merge
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
## Thinking Path
> - Paperclip orchestrates AI agents for zero-human companies, and
humans onboard into companies via invite links.
> - Some invite types (`company_join` with
`requires_company_admin_approval`) require an admin to approve the join
request after the invitee submits it.
> - While the requester waits, the invitee sees the
`AwaitingJoinApprovalPanel` in `InviteLanding`, which describes where
the admin needs to go to approve the request.
> - That panel rendered the destination — "Company Settings → Access" —
as two clickable `<a href="/company/settings/access">` links, even
though the surrounding copy is plainly addressed to the admin ("Ask
**them** to visit ..."), not the requester.
> - First-time invitees naturally click the only underlined link on the
screen, are sent to `/company/settings/access`, hit the "No company
access" panel (they have no membership yet), and conclude the invite
flow is broken.
> - This PR removes the navigation by rendering both "Company Settings →
Access" references as plain styled text (`<p>` / `<span>`), so the
guidance stays visible but cannot be followed by the requester.
> - The benefit is that the post-submit invite experience matches the
copy's intent — guidance for the admin, not navigation for the
requester.
Fixes#6784.
## What Changed
- `ui/src/pages/InviteLanding.tsx` — In `AwaitingJoinApprovalPanel`,
replace the two `<a href={approvalUrl}>Company Settings → Access</a>`
elements with `<p>` and `<span>` containing the same text. Remove the
now-unused `approvalUrl` constant.
- `ui/src/pages/InviteLanding.test.tsx` — Update the existing "pending
approval page" test: it previously asserted two anchor tags pointing at
`/company/settings/access`; it now asserts **zero** anchors while the
text "Company Settings → Access" still appears twice (in the "Approval
page" box and inline in the "Ask them to visit ..." sentence). Renamed
the test description from "...linked access instructions" to
"...non-clickable access instructions" to reflect the contract.
## Verification
```
pnpm vitest run ui/src/pages/InviteLanding.test.tsx
```
Result: 8 / 8 pass, including the updated "shows the pending approval
page with the company icon and non-clickable access instructions" case.
Manual reproduction (master @ `242a2c2f`,
`deploymentMode=authenticated`, `bind=lan`, embedded Postgres):
1. As instance admin, generate a `company_join` invite that requires
admin approval.
2. In a fresh browser profile, open the invite link.
3. Fill **Create your account** and submit.
4. The "Request to join \<company\>" panel appears.
5. Hover the "Company Settings → Access" mentions — no underline, no
link cursor; clicking does nothing. The text is still readable and the
surrounding copy ("Ask them to visit ...") still conveys the instruction
to the requester.
Before / after screenshots: see issue #6784 — the "before" state lands
users on `/company/settings/access` which renders "No company access".
After this PR the guidance is informational only.
## Risks
Low. UI-only change confined to one function in `InviteLanding.tsx` plus
its matching test. No API contracts, routes, or data shapes are
modified. The removed `approvalUrl` constant was only referenced by the
two anchor elements.
## Model Used
- Anthropic Claude Opus 4.7 (`claude-opus-4-7`, 1M context, extended
thinking enabled).
- Tools: file editing, Bash, Playwright reproduction against a
self-hosted Paperclip instance running master @ `242a2c2f`, and the
Paperclip monorepo's own Vitest suite for verifying the test update.
## 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 (will attach in PR thread)
- [x] I have updated relevant documentation to reflect my changes (no
docs files needed updating)
- [x] I have considered and documented any risks above
- [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 server config loader reads `.paperclip/config.json` and feeds it
into the shared Paperclip config schema.
> - When a config file exists but cannot be parsed or fails schema
validation, Paperclip should not silently ignore it.
> - The current `readConfigFile()` catch block treats invalid files the
same as missing files, so startup falls back to defaults while the
banner can still point at the ignored config path.
> - This pull request keeps the missing-file fallback, but makes present
invalid config files fail with a path-specific error.
> - The benefit is safer startup behavior and a clear diagnostic that
points at the invalid config field.
## Linked Issues or Issue Description
Fixes#8908
## What Changed
- Changed `readConfigFile()` to return `null` only when the config file
is absent.
- Added explicit errors for unreadable/invalid JSON config files.
- Added explicit Zod validation errors that include the config path and
invalid field path without printing config contents.
- Added server tests for missing config, invalid JSON, schema validation
failure, and valid config parsing.
## Verification
- `pnpm exec vitest run server/src/__tests__/config-file.test.ts`
- `pnpm --filter @paperclipai/server typecheck`
## Risks
Low risk for valid configs and missing configs. This intentionally
changes behavior for present invalid config files from silent fallback
to startup failure, which is the issue being fixed.
> 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 repository file inspection, GitHub
CLI, 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
- [ ] 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
Fixes#4206
## Thinking Path
> - Paperclip orchestrates AI agents on issues with checkout/release
semantics for execution locks
> - `POST /api/issues/:id/release` clears checkout and execution locks
when a heartbeat ends without finishing the issue
> - `issues.release()` unconditionally set `status: "todo"`, undoing
terminal and waiting states (`done`, `cancelled`, `in_review`,
`blocked`) set during the session
> - Agents reported status drift after release (e.g. `in_review` →
`todo`, `done` → `todo`), forcing manual PATCH recovery and risking
silent stalls
> - This pull request gates the `todo` re-queue to `in_progress` issues
only and preserves all other statuses on release
> - The benefit is lock cleanup without destroying workflow state agents
already recorded
## Linked Issues or Issue Description
- Fixes#4206 — `issues.release()` must not downgrade terminal/waiting
statuses
- Related internal incident: AIT-114 status drift on terminal issue
release (AI Trading Council)
## What Changed
- `server/src/services/issues.ts` — `releaseStatus` is `todo` only when
`existing.status === "in_progress"`; otherwise preserves
`existing.status`
- `server/src/__tests__/issue-stale-execution-lock-routes.test.ts` —
regression tests: release preserves done, cancelled, in_review, blocked
keeps `done` and clears lock fields
- `server/package.json` — patch bump `0.3.1` → `0.3.2`
- `server/CHANGELOG.md` — documents the fix
## Verification
```sh
pnpm --filter @paperclipai/server test issue-stale-execution-lock-routes
```
- 7/7 tests pass (parametrized done, cancelled, in_review, blocked)
(includes new `preserves terminal status when releasing a done issue`
and existing `in_progress` → `todo` on release)
- CI: Build, Typecheck, serialized server suites, e2e, Canary Dry Run
green on latest head `f31b55f`
## Risks
Low risk. Behaviour change is intentional: non-`in_progress` releases no
longer force `todo`. Agents that relied on release to re-queue
`in_review`/`blocked` work must PATCH status explicitly (documented in
agent lifecycle guidance). Rollback: revert this commit and redeploy
`@paperclipai/server` 0.3.1.
## Model Used
Anthropic Claude Opus 4.6 (extended thinking mode) — 200K context
window, tool use enabled. Assisted implementation and PR packaging for
AI Trading Council upstream port from local hotfix.
## 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 (N/A)
- [x] I have updated relevant documentation to reflect my changes
(CHANGELOG)
- [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
(re-review requested on head `f31b55f`)
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: brandon <brandonburr@gmail.com>
Bumps [react-i18next](https://github.com/i18next/react-i18next) from
17.0.9 to 17.0.10.
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/i18next/react-i18next/blob/master/CHANGELOG.md">react-i18next's
changelog</a>.</em></p>
<blockquote>
<h2>17.0.10</h2>
<ul>
<li>fix(warnings): the <code>useTranslation</code> and
<code>Trans</code> "You will need to pass in an i18next
instance" warnings now match the <code>useSSR</code> wording,
mentioning the props/context alternatives and the most common
unexplained cause at scale: duplicate react-i18next copies in monorepo
setups. The <code>Trans</code> variant also referenced the internal
<code>i18nextReactModule</code> name; it now points to the public
<code>initReactI18next</code> API.</li>
<li>feat(warnings): development-only warning
(<code>SUSPENDED_WHILE_LOADING</code>, logged once) right before
<code>useTranslation</code> suspends while translations are loading.
With the default <code>useSuspense: true</code> and no
<code><Suspense></code> boundary this previously surfaced as a
blank screen or a cryptic React error; the warning now names both fixes
(add a <code><Suspense></code> boundary or set
<code>react.useSuspense: false</code>). No-op in production builds; the
<code>process.env.NODE_ENV</code> check is wrapped so runtimes without a
<code>process</code> global (raw ESM in the browser, some edge runtimes)
stay silent instead of throwing.</li>
<li>ci: weekly workflow typechecking the test suite against
<code>@types/react@next</code> / <code>@types/react-dom@next</code>, so
the next React major's type changes (like the React 18
<code>TFunctionResult</code>/children wave) surface before user
reports.</li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="3b71c2766c"><code>3b71c27</code></a>
17.0.10</li>
<li><a
href="57c3500e6a"><code>57c3500</code></a>
build</li>
<li><a
href="c62476f4d1"><code>c62476f</code></a>
chore: sync package-lock with i18next ^26.2.0 devDependency bump</li>
<li><a
href="0126bd1cad"><code>0126bd1</code></a>
improve instance warnings (monorepo hint) + dev-only suspense warning +
weekl...</li>
<li>See full diff in <a
href="https://github.com/i18next/react-i18next/compare/v17.0.9...v17.0.10">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>
Bumps [ws](https://github.com/websockets/ws) from 8.19.0 to 8.21.1.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/websockets/ws/releases">ws's
releases</a>.</em></p>
<blockquote>
<h2>8.21.1</h2>
<h1>Bug fixes</h1>
<ul>
<li>Empty fragments are now counted toward the limit (a2f4e7c0).</li>
<li>The default values of the <code>maxBufferedChunks</code> and
<code>maxFragments</code> options have
been reduced (f197ac65).</li>
</ul>
<h2>8.21.0</h2>
<h1>Features</h1>
<ul>
<li>Introduced the <code>maxBufferedChunks</code> and
<code>maxFragments</code> options (2b2abd45).</li>
</ul>
<h1>Bug fixes</h1>
<ul>
<li>Fixed a remote memory exhaustion DoS vulnerability (2b2abd45).</li>
</ul>
<p>A high volume of tiny fragments and data chunks could be sent by a
peer, using
modest network traffic, to crash a <code>ws</code> server or client due
to OOM.</p>
<pre lang="js"><code>import { WebSocket, WebSocketServer } from 'ws';
<p>const wss = new WebSocketServer({ port: 0 }, function () {
const data = Buffer.alloc(1);
const options = { fin: false };
const { port } = wss.address();
const ws = new WebSocket(<code>ws://localhost:${port}</code>);</p>
<p>ws.on('open', function () {
(function send() {
ws.send(data, options, function (err) {
if (err) return;
send();
});
})();
});</p>
<p>ws.on('error', console.error);
ws.on('close', function (code, reason) {
console.log(<code>client close - code: ${code} reason:
${reason.toString()}</code>);
});
});</p>
<p>wss.on('connection', function (ws) {
ws.on('error', console.error);
ws.on('close', function (code, reason) {
console.log(<code>server close - code: ${code} reason:
${reason.toString()}</code>);
});
});
</code></pre></p>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="ae1de54330"><code>ae1de54</code></a>
[dist] 8.21.1</li>
<li><a
href="8e9511b86b"><code>8e9511b</code></a>
[ci] Trust Coveralls Homebrew tap</li>
<li><a
href="f197ac6514"><code>f197ac6</code></a>
[fix] Lower default values of <code>maxBufferedChunks</code> and
<code>maxFragments</code></li>
<li><a
href="8df8265c2f"><code>8df8265</code></a>
[ci] Update actions/checkout action to v7</li>
<li><a
href="a2f4e7c046"><code>a2f4e7c</code></a>
[fix] Count empty fragments toward the limit (<a
href="https://redirect.github.com/websockets/ws/issues/2329">#2329</a>)</li>
<li><a
href="e79f912cb3"><code>e79f912</code></a>
[pkg] Approve install scripts for bufferutil and utf-8-validate</li>
<li><a
href="4ea355d6d3"><code>4ea355d</code></a>
[doc] Document 32-bit signed integer coercion for option values</li>
<li><a
href="2120f4c8c6"><code>2120f4c</code></a>
[example] Remove uuid dependency</li>
<li><a
href="4c534a6b8a"><code>4c534a6</code></a>
[security] Add latest vulnerability to SECURITY.md</li>
<li><a
href="bca91adf15"><code>bca91ad</code></a>
[dist] 8.21.0</li>
<li>Additional commits viewable in <a
href="https://github.com/websockets/ws/compare/8.19.0...8.21.1">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>
Bumps [@codemirror/language](https://github.com/codemirror/language)
from 6.12.3 to 6.12.4.
<details>
<summary>Commits</summary>
<ul>
<li>See full diff in <a
href="https://github.com/codemirror/language/commits">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.
> - The CLI is one external control-plane surface for scripts and
operators.
> - Board API keys are the headless credential path for
board-authenticated automation.
> - The CLI already exposes board token create/list/revoke commands.
> - The existing token tests covered the generic agent token lifecycle,
but did not directly cover the board token lifecycle.
> - This pull request adds focused tests for board token creation,
listing, revocation, and expiration payload handling.
> - The benefit is safer CLI credential-management work without changing
runtime behavior.
## Linked Issues or Issue Description
No public issue found. This is a test coverage follow-up for CLI board
token lifecycle behavior described in
`doc/plans/2026-05-23-cli-api-parity.md`.
Related context: #4220 tracks/contains board API key product work; this
PR only adds tests for the current CLI command behavior on this branch.
## What Changed
- Added `token board create` coverage for `--ttl-days` expiration
payloads.
- Added `token board create` coverage for `--never-expires` payloads.
- Added `token board list` and `token board revoke` coverage, including
the DELETE route assertion.
## Verification
- `./node_modules/.bin/vitest run cli/src/__tests__/token.test.ts
--config cli/vitest.config.ts` passes: 1 file, 6 tests.
## Risks
Low risk. This is test-only coverage for existing CLI behavior.
> 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 GPT-5 Codex in Codex desktop, with repository inspection, shell
command execution, GitHub CLI usage, and code editing tools.
## 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: 馨冉 <xinxincui239@gmail.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The CLI is one external control-plane entry point for scripts,
operators, and agent handoff flows.
> - Prompt handoff intentionally maps prompts back to Paperclip work
objects: issues, comments, and optional wakeups.
> - The existing prompt tests covered the agent-authenticated path, but
the board-authenticated path had less direct coverage.
> - That left regressions in persona validation, board issue creation,
comment append, and no-wake behavior harder to catch.
> - This pull request adds focused tests for those board prompt handoff
paths.
> - The benefit is safer CLI parity work without changing runtime
behavior.
## Linked Issues or Issue Description
No public issue found. This is a test coverage improvement for the CLI
prompt handoff workflow described in
`doc/plans/2026-05-23-cli-api-parity.md`.
## What Changed
- Added `runBoardPrompt` coverage for rejecting agent persona profiles.
- Added board-authenticated issue creation coverage, including target
agent resolution and wakeup.
- Added board-authenticated comment append coverage with `wake: false`
to verify no wakeup request is sent.
## Verification
- `./node_modules/.bin/vitest run cli/src/__tests__/prompt.test.ts
--config cli/vitest.config.ts` passes: 1 file, 6 tests.
- `pnpm --filter paperclipai typecheck` was attempted, but this local
checkout fails before reaching this change because
`@paperclipai/plugin-sdk` dist artifacts are missing for server imports;
representative errors include `Cannot find module
@paperclipai/plugin-sdk` from `server/src/app.ts` and
`server/src/routes/plugins.ts`.
## Risks
Low risk. This is test-only coverage for existing CLI behavior.
> 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 GPT-5 Codex in Codex desktop, with repository inspection, shell
command execution, and code editing tools.
## 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
Co-authored-by: 馨冉 <xinxincui239@gmail.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The CLI is one of the main operator and agent-facing control
surfaces.
> - CLI commands build API paths from dynamic company, issue, project,
agent, and other resource identifiers.
> - Dynamic path segments need to be encoded so reserved characters
cannot reshape the request URL.
> - Empty dynamic path segments should fail locally instead of creating
malformed API routes.
> - The shared `apiPath` helper already implements those safeguards, but
its behavior was not directly covered in the common CLI tests.
> - This pull request adds focused coverage for path segment encoding
and empty-segment rejection.
> - The benefit is stronger regression coverage around a small but
security-relevant CLI routing helper.
## Linked Issues or Issue Description
- Bug: `apiPath` is the shared CLI helper for constructing API paths
with dynamic identifiers, but the common CLI tests did not directly
assert that dynamic segments are URL-encoded or that empty segments are
rejected before a request is made.
## What Changed
- Imported `apiPath` into `cli/src/__tests__/common.test.ts`.
- Added coverage that verifies reserved characters in dynamic path
segments are encoded.
- Added coverage that verifies empty and undefined dynamic path segments
throw before producing a malformed path.
## Verification
- `./node_modules/.bin/vitest run cli/src/__tests__/common.test.ts
--config cli/vitest.config.ts` passed (10 tests).
- `git diff --check` passed.
## Risks
- Low risk. This is test-only coverage for existing helper behavior.
- If future code intentionally wants query-string construction through
this helper, it should use static template text for the query string and
keep dynamic values as path segments or use a dedicated query builder.
> 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 GPT-5 via Codex, with code editing 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
- [ ] 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: 馨冉 <xinxincui239@gmail.com>
## Thinking Path
> - Paperclip is the open source control plane people use to manage AI
agents and their work
> - Operators rely on task detail properties and header actions to scan
linked resources and make quick decisions
> - External GitHub objects used a long pull-request label and repeated
mention counts that added visual noise without adding state
> - The task-detail star action also rendered as a labelled outline
button, unlike the compact star controls used elsewhere
> - These inconsistencies made dense task-detail surfaces slower to scan
and broke Paperclip's content-first visual language
> - This pull request shortens the GitHub pull-request label, removes
duplicate mention-count decoration, and aligns the detail star action
with the icon-only control pattern
> - The benefit is a calmer, more consistent task-detail experience with
accessible labels preserved for assistive technology
## Linked Issues or Issue Description
No matching public GitHub issue or open pull request was found.
**What happened?**
On the task detail surface, linked GitHub pull requests were labelled
`Github Pull Request` and could show a repeated `×N` mention count. The
detail-header star control used a labelled outline button rather than
the compact icon-only star pattern.
**Expected behavior**
Linked pull requests should use the concise `Github PR` label without
duplicate mention-count decoration, and the detail star action should
render as an accessible icon-only ghost button consistent with
neighboring controls.
**Steps to reproduce**
1. Open a task with a linked GitHub pull request mentioned more than
once.
2. Inspect the external-object property label and value row.
3. Inspect the star action in the task detail header.
**Paperclip version or commit:** `230126d80b` (`master` at preparation
time)
**Deployment / installation:** Local development, built from source.
**Scope:** Core UI; not adapter-specific, database-related, or
configuration-related.
## What Changed
- Render GitHub pull-request property labels as `Github PR`.
- Remove repeated external-object mention-count decoration from property
values.
- Render detail-header star controls as icon-only ghost buttons while
preserving `aria-label`, pressed, busy, error, and tooltip states.
- Expand component coverage for concise labels, duplicate-count
suppression, visual variants, and accessible star actions.
## Verification
- `pnpm --filter @paperclipai/ui exec vitest run
src/components/IssueProperties.test.tsx
src/components/StarToggle.test.tsx` — 54 tests passed.
- `pnpm --filter @paperclipai/ui typecheck` — passed.
- Visual review (before/after plus normal, starred, pending, and error
states):
45cb492185/star-toggle-review.html
- `pnpm check:token-gates` — reports five existing `#9627` violations in
files unchanged by this PR; the same values are present on `master`.
## Risks
- Low risk: changes are limited to task-detail presentation and tests.
- The star action remains fully accessible through its existing ARIA
label and tooltip, but it no longer displays visible text.
- External-object mention counts remain available in data; only the
redundant property-row decoration is removed.
> 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 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
- [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 control plane operators use to coordinate AI-agent
companies and review work needing attention.
> - The Inbox is the operator-facing surface that aggregates tasks
requiring attention across server state and shared client polling.
> - Archiving a task optimistically removed it, but ordinary background
activity and stale polling responses could make it reappear seconds
later.
> - The server therefore needs to distinguish genuine user-attention
events from routine agent/system activity.
> - The client also needs a bounded local archive guard across every
Inbox query path while the server mutation and in-flight polls settle.
> - This pull request fixes both resurrection paths and adds
race-focused regression coverage.
> - The benefit is stable archive behavior without hiding a genuine
archive failure after reconciliation or reload.
## Linked Issues or Issue Description
### Pre-submission checklist
- [x] Searched existing open and closed issues and pull requests; no
duplicate implementation was found.
- [x] Reproduced on `master` before this branch.
- [x] Confirmed this is a Paperclip core bug, not adapter or provider
behavior.
### What happened?
Archiving an Inbox task hid it optimistically, then background refresh
activity could insert it back into the list seconds later.
### Expected behavior
A successfully archived task remains hidden during normal polling. A
genuine failed archive may become visible again after reconciliation or
reload.
### Steps to reproduce
1. Open Inbox with a visible task.
2. Archive the task.
3. Wait for shared polling or routine agent activity to refresh task
data.
4. Observe the archived task reappear without a hard page refresh.
### Paperclip version or commit
`master` before this branch.
### Deployment mode
Built from source using the local development application.
### Installation method
Built from source (`pnpm`).
### Agent adapter(s) involved
Not adapter-specific; this is a core Inbox bug.
### Database mode
Not database-mode-specific.
### Access context
Board (human operator).
### Additional context
The failure had independent server and client causes: routine activity
could resurface archived rows server-side, while stale shared-poll
responses could bypass optimistic client removal.
## What Changed
- Restrict server-side Inbox resurfacing to explicit user-attention
events rather than any issue activity write.
- Add a bounded client-side archive guard with confirmation, failure
restoration, and cache reconciliation behavior.
- Apply the guard to Inbox rendering, badge counts, optimistic cache
updates, and shared-poll result application.
- Classify the generic compact Inbox query so stale shared-poll data
cannot bypass the guard.
- Add server visibility-matrix tests and UI race-condition regression
tests.
## Verification
- `pnpm --filter @paperclipai/ui exec vitest run
src/hooks/useSharedPolling.test.ts src/lib/inboxArchiveCache.test.ts
src/pages/Inbox.test.tsx` — 25 passed.
- `pnpm --filter @paperclipai/server exec vitest run
src/__tests__/issues-service.test.ts` — 107 passed.
- Branch rebased cleanly onto current `origin/master` before push.
## Risks
- Low-to-moderate behavioral risk: resurfacing is intentionally
narrower, so the server tests cover human comments, mentions,
interactions, and status transitions that must still regain attention.
- The client guard is bounded and cleared on mutation failure, limiting
the risk of hiding a task whose archive did not persist.
- No schema, migration, public API, workflow, dependency-lock, or
visual-token changes.
> 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
- Anthropic Claude via Claude Code (`claude_local`; prior
implementation/review run, exact underlying model ID and context window
were not retained in the handoff metadata), with repository tool use and
test execution.
- OpenAI `gpt-5.5` via Codex CLI for final review repair and PR
preparation, with reasoning, repository editing, GitHub tooling, 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/described the result 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 task identifier
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation where needed; no
documentation change is required for this bug fix
- [x] I have considered and documented risks above
- [x] All Paperclip-authored commits include the required co-author
trailer
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The Secrets UI lets operators navigate slash-delimited secret names
as folders
> - PR #9913 shipped create-in-folder behavior for company and per-user
secrets
> - The production behavior is already on master, but several
create-in-folder interaction paths lack retained regression coverage
there
> - Without that coverage, prefix composition, derived keys, prefix
removal, and staged empty folders could regress unnoticed
> - This pull request adds focused render tests without changing
production behavior
> - The benefit is safer maintenance of the folder-based secrets
workflow with a small, reviewable patch
## Linked Issues or Issue Description
- Refs #9913
## What Changed
- Added render coverage for creating a company secret from a folder and
deriving its key from the full slash-delimited name.
- Added coverage for per-user secret prefixes and exposing the full name
when the prefix chip is removed.
- Added coverage for inline folder-name validation and URL-backed
staging of an empty folder.
## Verification
- `env -u PAPERCLIP_IN_WORKTREE -u PAPERCLIP_WORKTREE_NAME -u
PAPERCLIP_CONFIG -u PAPERCLIP_HOME -u PAPERCLIP_INSTANCE_ID -u
PAPERCLIP_CONTEXT pnpm exec vitest run
ui/src/pages/Secrets.render.test.tsx` — 29 tests passed.
- `pnpm check:token-gates` — reports five existing `#9627` literals in
unrelated files; this PR changes no token-gated component code and
introduces no new violation.
## Risks
- Low risk: test-only change with no production, API, schema, migration,
dependency, or runtime behavior changes.
- The tests exercise existing DOM interactions and may need updates if
the Secrets creation UI copy or controls intentionally 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 CLI managed coding agent with repository inspection, code
execution, Git/GitHub, and Paperclip API tools. The managed harness does
not expose the exact underlying model ID or 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)
- [ ] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
— the assigned execution branch name is fixed by the task
- [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
documentation change is needed for test-only coverage
- [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 README is the first project surface most prospective users and
contributors see
> - The existing Star History presentation only showed one static chart
at the bottom of the page
> - The project also lacked the compact Star History Rank badge
alongside its other top-level badges
> - The chart should respect the reader's light or dark color scheme
while preserving the existing history link
> - This pull request adds the rank badge and replaces the chart with
responsive light/dark sources
> - The benefit is a more useful, theme-aware project-growth snapshot in
both prominent README locations
## Linked Issues or Issue Description
No public GitHub issue exists for this documentation request.
- **Problem / motivation:** The README's Star History presentation is
less visible than the other project badges and its chart does not adapt
to light and dark themes.
- **Proposed solution:** Add Star History's rank badge to the centered
badge row and use the provider's responsive `<picture>` markup for the
chart section.
- **Alternatives considered:** Keeping the current single static chart
would avoid markup changes but would not provide theme-aware rendering
or the requested rank badge.
- **Roadmap alignment:** Documentation-only presentation change; it does
not add or duplicate a product roadmap item.
- **Related PR:** #9922 refreshed the surrounding README roadmap content
before this focused follow-up.
## What Changed
- Added the Star History Rank badge to the README's centered badge row.
- Replaced the single Star History image with light- and dark-theme
chart sources using the supplied sealed chart URL.
- Preserved the existing Star History destination and accessible chart
label.
## Verification
- `git diff --check`
- Confirmed the badge, light chart, and dark chart endpoints each return
HTTP 200.
- Rendered the updated section through GitHub's GFM API and confirmed
GitHub preserves the linked `<picture>` and both theme sources.
- Confirmed the diff contains only `README.md`.
## Risks
- Low risk: documentation-only change with no runtime, API, dependency,
or migration impact.
- The chart depends on Star History's hosted badge/chart endpoints and
supplied sealed token remaining available.
> 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 `gpt-5.4` via Codex CLI; context-window size not exposed by the
runtime; medium reasoning with repository, shell, GitHub API, and live
HTTP verification tools.
## 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 telemetry subsystem (`packages/shared/src/telemetry`) flushes
event batches to an ingest endpoint; today the client drops batches
silently on 429 / 413 / 400 / network error with no re-queue and no
retry
> - The current behaviour is uncharacterised — there are no regression
tests pinning it, so any future retry work could accidentally change the
drop semantics without a test failing
> - Before adding retry logic it is critical to lock the current
baseline as an explicit, documented contract so regressions are
immediately visible
> - Additionally, the retry/cap design needs a config surface
(`maxEventsPerBatch`, `maxBodyBytes`, `maxPendingRetryBatches`,
`backoff`) so operators can tune behaviour without forking; that surface
should be defined and defaulted before the first consumer lands, not
after
> - This pull request adds characterisation tests that pin today's
silent-drop baseline and adds an optional, additive config surface with
resolved defaults — `client.ts` is untouched so no behaviour changes
> - The benefit is that the retry work in the follow-up PR can build on
a verified baseline, safe defaults, and a pre-wired config contract
rather than specifying everything at once
## Linked Issues or Issue Description
No pre-existing public GitHub issue. Underlying problem described below
(feature-request template).
**Problem or motivation**
The telemetry client in `packages/shared/src/telemetry/client.ts` has no
regression tests covering its error-path semantics. On 429 / 413 / 400 /
network failure the current code drains the whole queue via `splice(0)`
and discards the batch — silent drop, no retry. That behaviour is
correct for a best-effort client today, but nothing verifies it. When
retry logic lands, there is no safety net to confirm it does not
accidentally re-queue already-dropped batches. In parallel, the retry
design requires a typed config surface (`maxEventsPerBatch`,
`maxBodyBytes`, `maxPendingRetryBatches`, `backoff`) that the client and
its callers need to agree on before the first consumer is wired up.
**Proposed solution**
Two-phase additive scaffolding:
1. Lock the silent-drop baseline as pinned tests (temporary; Impl-2
replaces them when retry lands).
2. Add the config surface and defaults now so Impl-2 can consume them
immediately.
**Alternatives considered**
Writing the tests and config surface inside the retry PR — rejected
because it makes the retry diff much harder to review and loses the
regression-safety benefit of a committed baseline.
**Roadmap alignment**
Maintenance / client-correctness improvement; not a core roadmap
feature.
## What Changed
- **`client.test.ts`** — 5 new Phase-1 characterisation tests pinning
silent-drop on 429, 413, 400, and network-error. A
`stubFetchStatus(status)` helper DRYs the four drop cases. These are
temporary regression anchors: marked `TODO(impl-2): replace when retry
lands`.
- **`config.ts`** — `TELEMETRY_DEFAULTS` constant (frozen) +
`resolveCaps()` + `resolveTelemetryConfig()`. Provides a single exported
source of truth for defaults; nothing reads these yet.
- **`types.ts`** — 4 optional, additive `TelemetryConfig` fields:
`maxEventsPerBatch` (default 50), `maxBodyBytes` (default 524 288),
`maxPendingRetryBatches` (default 20), `backoff` (exp-with-jitter
shape). All optional and backward-compatible.
- **`index.ts`** — re-exports `TELEMETRY_DEFAULTS` and
`resolveTelemetryConfig` from the public package surface.
- **`client.ts`** — no changes (verified via `git diff --stat`).
## Verification
```bash
# Unit tests — 16 passed (5 new Phase-1 pins + 2 new Phase-2 config tests + 9 existing)
pnpm --filter @paperclipai/shared exec vitest run src/telemetry
# Type check — clean
pnpm --filter @paperclipai/shared typecheck
# Confirm client.ts untouched (zero output expected)
git diff HEAD~1 -- packages/shared/src/telemetry/client.ts
```
All three commands verified locally before push.
## Risks
**Low risk.** `client.ts` is untouched — no behaviour change. All four
modified files are additive:
- New tests cannot regress production behaviour.
- New config fields are `optional` and their defaults match the current
implicit constants in `client.ts`, so any future reader gets identical
semantics until it overrides them.
- `resolveTelemetryConfig` replaces no existing function; it is new.
- No envelope / wire change; no new PII / crypto / sink.
The `TODO(impl-2)` markers on the Phase-1 pins make their temporary
nature explicit in code.
## Model Used
- **Provider:** Anthropic
- **Model:** Claude Sonnet 4.6 (`claude-sonnet-4-6`)
- **Context window:** 200k tokens
- **Capabilities:** Tool use, code execution, agentic task completion
via Paperclip agent SDK
- **Mode:** Standard (no extended thinking)
## 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: Harold Kim <harold@paperclip.ing>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The CLI API client is the shared request layer used by command tests
and board/agent workflows.
> - Its unit tests replace the global `fetch` function so requests can
be asserted without a live server.
> - `vi.restoreAllMocks()` restores spies and mocks, but it does not
undo `vi.stubGlobal()` replacements.
> - That means a mocked global `fetch` can leak into later tests that
share the same Vitest worker.
> - This pull request makes the API client test cleanup match the safer
CLI test pattern by unstubbing globals after each test.
> - The benefit is more reliable CLI test isolation without changing
runtime behavior.
## Linked Issues or Issue Description
- Bug: `cli/src/__tests__/http.test.ts` stubs global `fetch` in multiple
tests but only calls `vi.restoreAllMocks()` during cleanup. Vitest does
not use `restoreAllMocks()` to undo `vi.stubGlobal()`, so later tests in
the same worker can inherit a mocked `fetch` and exercise the wrong
behavior.
## What Changed
- Added `vi.unstubAllGlobals()` to the API client test `afterEach`
cleanup.
- Kept the change limited to test isolation; no runtime code changed.
## Verification
- `./node_modules/.bin/vitest run cli/src/__tests__/http.test.ts
--config cli/vitest.config.ts` passed (5 tests).
- `git diff --check` passed.
## Risks
- Low risk. This only changes test cleanup and should make tests less
order-dependent.
- If a future test intentionally relies on a global stub persisting
across test cases, it will need to move that setup into its own
`beforeEach`; that would be a healthier test shape.
> 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 GPT-5 via Codex, with code editing 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
- [ ] 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: 馨冉 <xinxincui239@gmail.com>
## 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 control plane people use to manage AI-agent
companies.
> - Operators need to grant secrets to specific agents safely and
efficiently.
> - The secrets access sheet previously used a native select while
similar agent-assignment surfaces use a searchable picker.
> - That inconsistency makes finding an agent slow and error-prone as
companies grow.
> - This pull request reuses the shared agent picker behavior for
single-agent secret access grants.
> - The benefit is a consistent, searchable selection experience without
changing secret-access semantics.
## Linked Issues or Issue Description
Refs #9797
**Problem / motivation:** The in-sheet agent access form introduced in
#9797 renders every grantable agent in a native select. In companies
with many agents, operators cannot filter by name or title and the
experience differs from other agent-selection surfaces.
**Proposed solution:** Add a single-select variant beside the existing
`AgentMultiSelect`, then use it in the secret access grant form.
**Alternatives considered:** Keeping a native select would preserve less
code but would not scale or align with existing searchable agent
selection.
**Roadmap alignment:** This is a focused usability fix for an existing
core control-plane surface and does not overlap an unstarted roadmap
initiative.
## What Changed
- Added `AgentSelect`, a searchable single-agent popover that filters by
agent name and title.
- Replaced the secret access form's native select with the shared
searchable picker.
- Added focused coverage for filtering, selecting, callback behavior,
and popover closure.
## Verification
- `pnpm exec vitest run ui/src/components/AgentMultiSelect.test.tsx` —
passes (3 tests).
- `pnpm --filter @paperclipai/ui typecheck` — passes.
- `pnpm check:token-gates` — branch adds no violations; the command
currently fails on five unchanged `#9627` literals already present on
`master`.
- Manual: open Secrets, choose a secret, add agent access, filter by
agent name or title, select the result, and grant access.
## Risks
- Low risk: the change is limited to agent selection UI and preserves
the existing grant request payload.
- The new picker depends on the existing popover/input primitives and
resets its filter when closed.
> 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-series coding agent (exact runtime model ID and
context-window size are not exposed), with reasoning, repository tool
use, terminal execution, and GitHub CLI 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
- [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 control plane people use to organize
and govern companies of AI agents
> - The README is the project’s front door, while `ROADMAP.md` is the
detailed statement of shipped and planned product direction
> - The existing README roadmap had fallen behind capabilities that are
already available and did not explain the product’s organizing model
> - That drift made current features look unfinished and left the
relationship between tasks, organization, training, and infrastructure
implicit
> - This pull request synchronizes the roadmap preview with the detailed
roadmap and introduces the four-pillar product framing
> - The benefit is a more accurate, coherent first impression for users
and contributors, with theme-aware visuals that render correctly on
GitHub
## Linked Issues or Issue Description
- **Subsystem affected:** Documentation (`README.md`, `ROADMAP.md`, and
documentation assets); no runtime subsystem changes.
- **Problem or motivation:** The README roadmap understated shipped
capabilities, the detailed roadmap had corresponding stale statuses, and
the README did not clearly explain the four product pillars that connect
Paperclip’s task, organization, training, and infrastructure surfaces.
- **Proposed solution:** Refresh both roadmap views from one ordered
list, add the four-pillars section and light/dark graphic, and keep
future capabilities explicitly marked as planned or partial.
- **Alternatives considered:** Updating only `ROADMAP.md` would leave
the project’s front page stale; updating only the README would create a
second source of truth; a single non-theme-aware image would be less
legible for half of GitHub’s color schemes.
- **Roadmap alignment:** This change directly updates `ROADMAP.md` and
its README preview rather than introducing an unplanned runtime feature.
- **Additional context:** Searched open pull requests for related README
roadmap/four-pillars work and found no duplicate.
## What Changed
- Added a **The four pillars** section covering Agentic Task Manager,
Org Chart for Agents, Agent Employee Training, and Agentic OS, with a
theme-aware `<picture>` graphic.
- Flipped four roadmap entries from ⚪ to ✅: Cloud / Sandbox agents,
Artifacts & Work Products, Deep Planning, and Enforced Outcomes.
- Added shipped ✅ entries for MCP Tool Gateway & Apps, per-agent Secrets
Manager access, activity logging and action attribution, self-healing
recovery, and agent evals/feedback; expanded the skills entry to include
Skill Studio and Skills Store.
- Added planned ⚪ entries for Bring-your-own-ticket-system and Connected
Apps, resolving the README FAQ inconsistency around external ticket
systems.
- Moved Cloud deployments to 🟡 and documented the shipped multi-tenant
isolation, local-to-cloud sync, and cloud-managed bootstrap scope.
- Synchronized the README preview and `ROADMAP.md` ordering/statuses: 18
✅, 9 ⚪, and 1 🟡 entry.
## Verification
- `git diff --check origin/master...HEAD`
- Focused Node consistency check confirms README and `ROADMAP.md` have
the same 28 ordered entries/statuses and expected 18 ✅ / 9 ⚪ / 1 🟡
counts.
- `file doc/assets/four-pillars-light.png
doc/assets/four-pillars-dark.png` confirms both assets are valid
3840×2160 RGB PNGs.
- Visual QA rendered the README through GitHub’s GFM API in Chromium at
GitHub content width under both light and dark `prefers-color-scheme`
modes; the correct graphic swapped per theme, all links resolved, the
section placement/table were correct, and `ROADMAP.md` matched the
README preview.
### Visual Evidence
**Light theme**
**Dark theme**
- Prettier was not run because this repository does not expose a
configured `prettier` binary; markdown whitespace was checked with `git
diff --check` and the rendered QA above.
## Risks
- Low risk: documentation and image assets only; no application code,
API contract, schema, dependencies, lockfile, or workflow changes.
- Roadmap status wording can become stale as product work evolves, so
future shipped capabilities should update both the README preview and
`ROADMAP.md` together.
> 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, exact model ID `gpt-5.6-sol`, with reasoning,
terminal/tool use, GitHub API access, and Paperclip control-plane
integration. The model context-window size was not exposed by this
execution 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 issue-comments API (`POST /api/issues/:id/comments`) attributes
each comment to the run that created it via `created_by_run_id`, a
foreign key into `heartbeat_runs`
> - In multi-agent local control-plane usage, board/session clients
sometimes forward an `X-Paperclip-Run-Id` that is not a real run row — a
non-UUID client request id, a synthetic string, or a since-deleted run
> - That value was written straight to the FK column, so the insert died
with a Postgres FK violation and the endpoint returned HTTP 500,
breaking agent coordination
> - This PR resolves the run id defensively before insert: reject
non-UUID shapes, verify the row exists for the company, and null out
anything unresolvable while logging a warning
> - The benefit is that a bad run-id header degrades gracefully to an
unattributed comment (201) instead of a 500, so comment creation stays
up
## Linked Issues or Issue Description
No public issue exists; describing inline (bug):
**What happened:** `POST /api/issues/:id/comments` returns HTTP 500 when
the request carries an `X-Paperclip-Run-Id` that does not correspond to
a row in `heartbeat_runs` (non-UUID value, synthetic client id, or
deleted run). The value is written to the `created_by_run_id` FK, and
Postgres rejects the insert with a foreign-key violation (SQLSTATE
23503).
**Expected:** the comment is created (HTTP 201); an unresolvable run id
is dropped to `null` rather than failing the request.
**Impact:** in multi-agent usage, comment creation — and the agent
coordination that depends on it — fails whenever a client forwards a run
id that isn't a live run.
## What Changed
- Add `resolveCommentCreatedByRunId(dbOrTx, companyId, runId)` — trims
and validates UUID shape, then checks existence in `heartbeat_runs`
scoped to the company; returns `null` for missing/invalid ids.
- `addComment` now resolves the run id through that helper before insert
and logs a warning when a supplied run id is dropped.
- Add embedded-Postgres regression tests for the three cases (non-UUID
header, unknown UUID, valid run id).
## Verification
- `pnpm --filter server test issues-service` — the new
`issueService.addComment createdByRunId` block passes.
- Cases covered: non-UUID header → 201, `createdByRunId: null`; UUID
absent from `heartbeat_runs` → 201, `null`; valid run id present for the
company → preserved.
## Risks
Low. Purely defensive — valid run ids are still preserved, only
unresolvable ones are nulled. Adds one indexed, tenant-scoped `SELECT`
per comment insert.
## Model Used
Claude Opus 4.8 (extended thinking), via the Paperclip PR-triage
cockpit, produced the added regression tests and this description. The
original implementation is by @digitalflanker-ux; the author's model is
unspecified.
## 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)
- [ ] 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 (related: #4795 same fix; #8065 sibling FK-guard on the
activity-log path)
- [x] I have either (a) linked existing issues OR (b) described the
issue in-PR following the relevant issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links
- [ ] My branch name describes the change and contains no internal
Paperclip ticket id
- [ ] 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: Claude Haiku 4.5 <noreply@anthropic.com>
Co-authored-by: Andrew Aymeloglu <aaymeloglu@gmail.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Operators rely on Paperclip server logs for maintenance, incident
triage, and support handoffs.
> - The HTTP logger persisted request metadata and only redacted
authorization headers.
> - Request cookies and set-cookie headers can contain active session
material and should not be written to durable logs.
> - This pull request keeps the fix intentionally narrow: centralize the
HTTP log redaction path list and include cookie-bearing headers.
> - The benefit is lower credential/session leakage risk from routine
server.log collection or sharing.
## Linked Issues or Issue Description
No GitHub issue exists for this exact local finding. Inline bug report:
- Type: security/privacy bug.
- Affected area: server HTTP logging middleware.
- Observed problem: local Paperclip maintenance found raw cookies
present in server.log.
- Expected behavior: durable HTTP logs redact authorization and
cookie-bearing request/response headers.
- Impact: anyone with access to copied/exported logs could see
session-bearing cookie values.
- Related/open PRs found during dedup search: #7242, #7306, #7346. This
PR is the minimal local fix branch created from the verified local
maintenance patch; those PRs may be better upstream candidates if
maintainers prefer their broader coverage.
## What Changed
- Added `HTTP_LOG_REDACT_PATHS` for HTTP logger redaction paths.
- Kept existing `req.headers.authorization` redaction.
- Added redaction for `req.headers.cookie`, request `set-cookie`, and
response `set-cookie` paths.
- Added focused tests asserting the required redaction paths are present
and that pino-http output redacts live request/response header secrets.
## Verification
- `pnpm exec vitest run server/src/__tests__/http-log-redaction.test.ts`
- `pnpm --filter @paperclipai/server typecheck`
- Pre-commit TruffleHog scan: 0 verified/unverified secrets.
- PR CI observed passing so far for policy, Typecheck + Release
Registry, Build, e2e, Socket, Snyk, security-review, and
serialized/workspace suites; remaining jobs may still be running.
## Risks
- Low runtime risk: this only expands pino redaction paths.
- Possible coverage risk: broader redaction helpers in related PRs may
cover more serialized variants beyond the pino-http request/response
header pipeline tested here.
- No migrations, schema changes, or UI changes.
## Model Used
- OpenAI Codex via Hermes Agent, model gpt-5.5, tool-using coding/ops
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 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
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Andrew Aymeloglu <aaymeloglu@gmail.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Operators store company and user secrets under human-readable names
> - Slash-delimited names already express useful hierarchy, but the
Secrets page previously rendered them as one flat list
> - Large secret collections therefore became harder to scan, navigate,
and create within consistently named groups
> - A client-derived folder model preserves the existing server contract
while making those names navigable
> - This pull request adds folder browsing, URL-addressable paths,
global search, and create-in-folder behavior without adding folder
records
> - The benefit is a more scalable secrets workflow with no migration or
API compatibility risk
## Linked Issues or Issue Description
Slash-delimited secret names such as `dev/github/oauth/clientid`
currently appear as raw flat rows. This change treats name prefixes as
client-side navigation folders on the main Secrets tab while leaving
stored names, API contracts, validation, and the database unchanged.
- Wireframes and interaction specification:
https://pages.paperclip.ing/pap-14698-secrets-folders/
- Folder paths are derived only from secret names; there is no
server-side folder entity or data-model change.
## What Changed
- Added pure secret-path utilities for normalized segments, breadcrumbs,
nested folder listings, counts, and leaf/path rendering.
- Added a Folders/Flat view to the main Secrets tab with folder-first
sorting, breadcrumbs, empty-folder states, filters, and global search
results.
- Added URL navigation through `?path=<normalized/path>` so deep links,
reload, browser Back, and new-tab folder navigation work.
- Persisted the preferred view in `localStorage` under
`paperclip.secrets.viewMode`; an explicit `?path=` deep link takes
precedence for that visit.
- Added create-in-folder behavior with a removable prefix chip, staged
New folder paths, inline segment validation, and full-name key
derivation.
- Kept My secrets flat while rendering slash-delimited names with muted
paths and emphasized leaves.
- Added unit/render coverage for path helpers, folder navigation, and
create-in-folder behavior.
## Verification
- `pnpm exec vitest run ui/src/pages/secrets/secret-path.test.ts
ui/src/pages/Secrets.render.test.tsx` — 35 tests passed.
- `pnpm -r typecheck` — passed.
- `pnpm test:run` — server suite passed 2,707 tests and UI suite passed
3,022 tests; one unrelated CLI doctor test was environment-sensitive
because this agent inherited AWS variables.
- `env -u AWS_PROFILE -u AWS_ACCESS_KEY_ID -u AWS_SECRET_ACCESS_KEY -u
AWS_SESSION_TOKEN -u AWS_REGION -u AWS_DEFAULT_REGION pnpm exec vitest
run cli/src/__tests__/secrets.test.ts` — 8 tests passed.
- `pnpm build` — passed.
- `pnpm check:token-gates` — feature files are clean; the command
reports five existing `#9627` color literals outside this diff.
## Risks
- Low data risk: folders are derived client-side from existing names,
with no schema, migration, API, or stored-name changes.
- URL behavior changes only the main Secrets tab and uses the additive
`?path=` contract.
- View preference is browser-local and scoped to the
`paperclip.secrets.viewMode` key.
- Renaming or deleting the last secret under an open prefix
intentionally leaves the user on an empty-folder state instead of
redirecting.
- “Move to folder…” bulk prefix rename remains deferred because it is a
multi-secret mutation with separate conflict and partial-failure
semantics.
> 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 `gpt-5.5` via Codex CLI for PR preparation, verification,
GitHub/Paperclip tool use, and codebase analysis; the managed harness
does not expose the active context-window size.
- Anthropic Claude Opus 4.8 (1M context) and Claude Fable 5 assisted
earlier implementation/design commits, as recorded in their commit
trailers; both used repository and code-editing tools.
## 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
— the assigned shared execution branch name is fixed for this work
- [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 Fable 5 <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
> - Paperclip instances can be self-hosted, or provisioned and managed
by a cloud control plane that authenticates users through trusted
headers validated against `PAPERCLIP_CLOUD_TENANT_SERVER_TOKEN`
(`resolveCloudTenantActor`)
> - In `authenticated` deployment mode, the health route reports
`bootstrapStatus: bootstrap_pending` until at least one `instance_admin`
exists, and the UI locks everyone out at the "waiting on its first
admin" claim screen until then — correct for self-hosted instances,
where a human operator must claim the instance
> - But the cloud-tenant trust middleware, by deliberate security
hardening, never grants `instance_admin` and actively purges legacy
grants — so a cloud-managed instance can never leave
`bootstrap_pending`: the gate demands a role the middleware forbids
> - Every control-plane-provisioned instance is therefore permanently
locked at the claim screen even though its users and memberships exist
> - This pull request makes the gate cloud-aware: when the tenant server
token is configured, the instance is considered bootstrapped, because
the control plane owns identity and there is no operator claim step
> - The benefit is that cloud-managed instances become usable while
self-hosted behavior stays byte-for-byte identical, now pinned by a
previously missing regression test
## Linked Issues or Issue Description
Refs #2927 (introduced the browser-native first-admin bootstrap flow
this gate feeds). No existing public issue for the deadlock; inline
description per the bug report template:
- **What happened?**: an instance configured with
`PAPERCLIP_DEPLOYMENT_MODE=authenticated` and
`PAPERCLIP_CLOUD_TENANT_SERVER_TOKEN` reports `bootstrapStatus:
bootstrap_pending` forever. All users — including ones created via the
trusted-header path with owner-level company membership — are locked out
at the "This Paperclip is waiting on its first admin" screen.
- **Expected behavior**: a control-plane-managed instance has no
first-admin claim step; users arriving with control-plane identity
should reach the app.
- **Steps to reproduce**:
1. Run the server with `PAPERCLIP_DEPLOYMENT_MODE=authenticated` and a
`PAPERCLIP_CLOUD_TENANT_SERVER_TOKEN` set
2. Create users only through trusted cloud headers (the middleware
upserts them but never grants `instance_admin`, and purges any legacy
grants)
3. `GET /api/health` → `bootstrapStatus` stays `bootstrap_pending`; the
UI shows the claim screen for every visitor, and no supported path
exists to create the `instance_admin` the gate requires
- **Paperclip version or commit**: reproducible on `master` as of
2026-07-20; present since the cloud-tenant `instance_admin` purge
hardening landed.
## What Changed
- `server/src/middleware/auth.ts`: new exported
`isCloudManagedInstance()` predicate beside the trust middleware that
defines the tenant-token contract.
- `server/src/routes/health.ts`: the authenticated-mode first-admin gate
is skipped when the instance is cloud-managed; `bootstrapStatus` reports
`ready`.
- `server/src/__tests__/health.test.ts`: two new tests — authenticated
without the token → `bootstrap_pending` (previously untested regression
baseline), and with the token → `ready` despite zero instance admins.
## Verification
- `pnpm vitest run src/__tests__/health.test.ts` in `server/` — 13/13
- `pnpm vitest run src/middleware/cloud-tenant-actor.test.ts` — 6/6
- Manual: with the env vars from the repro steps set, `GET /api/health`
now returns `bootstrapStatus: "ready"`; without the token, behavior is
unchanged
## Risks
- None for self-hosted deployments: without the env var the gate is the
prior behavior, now pinned by the new regression test.
- For cloud-managed instances the claim screen and
`bootstrapInviteActive` flow no longer appear — intended; browser-based
claim was already disabled in that configuration.
## Model Used
- Claude (Anthropic) — model id `claude-fable-5`, via the Claude Code
CLI harness with tool use (shell, file edits, test execution). Diagnosis
and change 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
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes (n/a —
behavior documented in code comments and pinned by tests)
- [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
> - 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
Release changelog for **v2026.720.0** (released 2026-07-20). Supersedes
#9612 (the 7/15 draft — that cut was never released, no `v2026.715.0`
tag exists).
Range is the full `v2026.707.0..master` — **195 commits from 16
contributors**.
## New since the 7/15 draft
- **Decision Training** — library + inspector, snapshot foundation,
decision image galleries
- **Built-in Summarizer & summary slots**
- **Skill organization** — nested folders, "My Skills",
import-from-project, open-by-default company skill policy (alongside
Skill Studio)
- **Issue monitor visibility** across task surfaces + issue properties
- **Reworked secrets dialog** with in-sheet per-agent access
- **Active PR-gardening workflow**
- Recovery / hot-restart / codex-auth reliability fixes and inbox/list
polish
No `BREAKING:` signals — DB migrations are additive and run
automatically (captured in the Upgrade Guide).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
## Thinking Path
> - Paperclip is the open source control plane people use to supervise
AI-agent companies.
> - Operators need task execution state to be visible where they read,
reply to, and manage an issue.
> - Scheduled monitor state was easy to miss because it lived in a
description-area card and used inconsistent, mostly static time copy.
> - The task page, composer, retry card, and properties panel therefore
needed one shared monitor-state and countdown language.
> - This pull request adds shared live-ticking time utilities, moves
monitor status into a top-of-page banner and composer strip, and
redesigns the properties row with complete read-only details.
> - The benefit is that operators can immediately understand when an
agent resumes, why it is waiting, and how to act without hunting across
the page.
## Linked Issues or Issue Description
Issue monitors can schedule a future agent check, retry, or wake, but
the UI did not present that state consistently or prominently. The
existing description-area activity card competed with issue content, the
composer did not explain that replying wakes the agent early, compact
properties copy truncated important details, and relative times did not
share a live two-unit formatter.
This change makes scheduled monitor state visible and consistent across
the issue header, reply composer, properties panel, and scheduled-retry
card. Related prior server-side recovery visibility work: #9629
(distinct scope).
## What Changed
- Added shared two-unit monitor ETA/offset formatters and a live-ticking
countdown hook, including compact absolute-time rules for Today,
weekday, and cross-year dates.
- Replaced the description-area monitor activity card with a top-of-page
status banner and added an inline composer strip that explains replies
wake the agent before the scheduled check.
- Redesigned the properties Monitor row as readable two-line copy with
attempt state, due/overdue/cleared wording, click-to-edit behavior, and
a hover/tap details tooltip.
- Adopted the shared two-unit formatting in the scheduled-retry card and
added focused coverage for monitor formatting, state transitions,
visibility, and controls.
- Added the approved wireframe package and published reference:
https://pages.paperclip.ing/pap-14557-monitor-visibility/
## Verification
- `pnpm vitest run ui/src/lib/issue-monitor.test.tsx
ui/src/components/IssueMonitorBanner.test.tsx
ui/src/components/IssueProperties.test.tsx` — 3 files, 74 tests passed.
- `pnpm -r typecheck` — passed.
- `pnpm build` — passed with existing Vite optimization/chunk-size
warnings.
- UX review approved the rendered real components across desktop/mobile,
light/dark, and the scheduled/retrying/due/overdue/cleared/none state
matrix.
- QA passed 6/6 criteria in Chromium, including a live countdown
transition without refresh. Review evidence included the P2 banner
state-matrix screenshot and the P3 properties-row and details-tooltip
screenshots, plus a dark-mode capture.
- `pnpm check:token-gates` currently reports five pre-existing `#9627`
comment literals on `origin/master`; this branch introduces none of
those literals or any new token violation.
## Risks
- Low-to-moderate UI behavior risk: monitor copy and placement change
across several issue surfaces, but all derive from one shared state
builder and focused tests cover the state matrix.
- Countdown rendering wakes once per minute while a visible monitor is
scheduled; the hook is limited to active monitor surfaces and stops when
hidden.
- No database, API contract, migration, telemetry, or authorization
behavior changes.
> 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-5.6-sol`; context-window metadata was unavailable
in this runtime; reasoning, repository tool use, command execution, and
test execution enabled. Earlier implementation commits were assisted by
Claude Fable 5 and Claude Opus 4.8 (1M context), as credited in their
commit trailers.
## 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: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source control plane people use to manage AI
agents and the credentials those agents need for work.
> - The secrets UI is responsible for making credential creation
understandable and for showing where credentials are used.
> - The create flow previously exposed an editable generated key too
early, used an unnatural field order, and rendered uneven provider tab
rows.
> - The detail sheet also lacked an in-context way to see and manage
which agents reference a selected secret.
> - This pull request makes the create dialog follow a predictable
name-to-value flow and adds agent access management directly to the
secret details sheet.
> - The benefit is a clearer secrets workflow with fewer accidental key
edits and less navigation when granting or revoking agent access.
## Linked Issues or Issue Description
**Subsystem affected:** `ui/` — React + Vite board UI
**Problem or motivation:** Creating a secret currently makes its
generated key look immediately editable, places fields in an awkward
keyboard order, and applies provider-tab overrides that produce uneven
rows. After creation, operators cannot inspect or change agent access
from the selected secret's detail sheet.
**Proposed solution:** Generate the key from the path-style name and
keep it read-only until an explicit Edit action; place Value directly
after Name; use the standard tab sizing; and add an Agent access section
that reads and updates `secret_ref` / `user_secret_ref` environment
bindings through agent adapter configuration.
**Alternatives considered:** Keeping access management only on agent
configuration screens was rejected because it hides a secret-centric
question—“which agents can use this?”—and requires repetitive
navigation. Keeping the key always editable was rejected because the
generated value should be the safe default.
**Roadmap alignment:** No matching item was found in `ROADMAP.md`; this
is a focused usability and access-management improvement to the existing
secrets surface.
**Additional context:** GitHub search found no duplicate PR for this
dialog and in-sheet access change. PR #9321 also mentions user-secret
resolution but addresses unrelated skills-route behavior.
## What Changed
- Auto-generate the create-secret key from Name, keep it read-only by
default, and expose an explicit Edit action.
- Use a path-style Name placeholder (`/dev/foo/bar`) and place Value
immediately after Name for natural keyboard navigation.
- Remove tab sizing/whitespace overrides that caused uneven provider-tab
row heights.
- Add an Agent access section to the Details tab that lists referencing
agents and grants or revokes `secret_ref` / `user_secret_ref`
environment bindings in place.
- Cover company secrets and each-user definitions with focused render
tests.
## Verification
- `pnpm exec vitest run ui/src/pages/Secrets.render.test.tsx` — 17/17
passed.
- `pnpm --filter @paperclipai/ui typecheck` — passed.
- `pnpm check:token-gates` — changed files are clean; the
repository-wide command currently reports five pre-existing `#9627`
violations in unrelated files (`Sidebar.tsx`, `Inbox.tsx`,
`IssueDetail.tsx`, `Issues.tsx`, and `Routines.tsx`).
- Additional secrets verification completed during implementation: 35/35
adjacent secrets tests passed, and the flow was checked in a real
browser in light and dark themes.
## Risks
- Agent access mutations update adapter environment configuration, so
malformed legacy env entries could affect how a binding is displayed or
revoked.
- Each-user definitions use `user_secret_ref` rather than `secret_ref`;
focused tests cover selecting the correct binding type.
- No database or API contract changes are included.
> 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
- Anthropic Claude Fable 5 assisted with the implementation using
repository-aware code editing and test execution.
- OpenAI GPT-5.5 via Codex CLI assisted with PR preparation, branch
hygiene, focused verification, GitHub operations, and review/check
loops.
## 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: Claude Fable 5 <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
> - Its web UI is a long-lived single-page app; users leave tabs open
for hours/days
> - We chased multi-GB tab memory growth through several fixes (#9569,
#9624, #9627, #9701) that cut real churn — but the footprint kept
climbing
> - A heap-snapshot diff on a 12-hour tab finally showed the true cause:
of 13.2M heap nodes, **12.1M were `PerformanceMeasure` objects** — React
19.2 emits a `performance.measure()` per component render for its
DevTools "Performance Tracks" and never clears them
> - These are native objects, so `performance.memory` never reported
them (it read a flat ~74 MB while the real heap was ~308 MB), which is
why our earlier heap/DOM sampling looked stable while the footprint
ballooned
> - This pull request periodically clears the User Timing measure
buffer, since nothing in the app consumes it
> - The benefit is that long-lived tabs stop accumulating millions of
native `PerformanceMeasure` objects, eliminating the remaining unbounded
growth
## Linked Issues or Issue Description
No public GitHub issue exists; describing inline per CONTRIBUTING.md →
"Link Issues or Describe Them In-PR", following the bug report template.
This is the root-cause fix for the memory growth chased in #9569 / #9624
/ #9627 / #9701.
**What happened?**
Long-lived browser tabs grew to multiple GB of memory footprint over
hours/days. A heap snapshot of a 12h tab showed **13.2M nodes, of which
12.1M were `PerformanceMeasure`** (vs ~1.0M total nodes on a fresh tab).
Live capture showed **~340 `performance.measure()` calls/sec**, named
after React components with a `detail.devtools` payload — React 19.2's
"Performance Tracks". Nothing ever clears them, so they accumulate
without bound.
**Expected behavior**
A long-lived tab should not accumulate millions of `PerformanceMeasure`
entries. Idle/long-running tabs should hold a bounded footprint.
**Steps to reproduce**
Open the app (React 19.2 production build), leave a tab open with normal
activity, and run `performance.getEntriesByType('measure').length`
periodically — it climbs unbounded (12.3M after ~12h).
`performance.clearMeasures()` drops it to near-zero and frees the
memory.
**Paperclip version or commit**
Branch `fix/react-perf-measure-leak`, off `master` (after #9701).
**Deployment mode**
Local dev (`pnpm dev` build served by the dev instance), web UI. React
19.2.7. Not adapter-specific.
## What Changed
- **`ui/src/lib/perf-measure-reaper.ts` (new)** —
`startPerfMeasureReaper(intervalMs = 10_000)` clears
`performance.clearMeasures()` on a timer and returns a stop function. It
never calls `getEntriesByType('measure')` (which would materialize the
huge buffer). Honors a `window.__paperclipKeepPerfMeasures = true`
opt-out so a developer recording a React Performance Track in DevTools
can keep the entries.
- **`ui/src/main.tsx`** — start the reaper at app boot.
- Only *measures* are cleared — React's tracks pass explicit start/end
times and leave no marks, and the app doesn't use the User Timing API at
all (verified), so nothing else is affected.
## Verification
- **Root cause proven** by heap-snapshot diff: fresh tab ~1.0M nodes vs
12h tab 13.2M nodes, 12.1M of them `PerformanceMeasure`; live capture
showed ~340 measures/sec with `detail.devtools` and React component
names.
- **Fix proven live**: `performance.clearMeasures()` on the aged tab
dropped the buffer from **12,418,266 → 1,500** and reclaimed the memory
(heap `perf.memory` 308 → 269 MB, plus the ~12M native objects, which
are the bulk of the footprint).
- `vitest`: `perf-measure-reaper.test.ts` — interval clearing, `stop()`,
opt-out flag, and no-API no-op. All pass.
- `tsc -b` clean.
## Risks
Very low, client-only.
- Clearing the User Timing measure buffer only affects the DevTools
Performance panel's React track *history*; normal users never consume
it. Developers who want to record it can set
`window.__paperclipKeepPerfMeasures = true`.
- Only `performance.clearMeasures()` is called (not `clearMarks`), so
any mark-based timing elsewhere is untouched; a grep confirmed the app
makes no `performance.mark()/measure()` calls of its own.
- Adds exactly one 10s interval (negligible), and `clearMeasures()` does
not materialize the buffer.
Note: this is a React 19.2 upstream behavior (its performance tracks are
emitted in the production build and never cleared). If React later gates
or clears them, this reaper can be removed.
## Model Used
- **Provider:** Anthropic, via the Claude Code CLI.
- **Model:** Claude Opus 4.8 (`claude-opus-4-8`).
- **Reasoning mode:** Extended thinking enabled.
- **Capabilities used:** tool use (shell, file editing), the Chrome
DevTools MCP to reproduce and profile, and a streaming heap-snapshot
parser to identify the `PerformanceMeasure` accumulation.
## 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 (continues #9701; no 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 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 considered and documented any risks above
- [ ] I have updated relevant documentation to reflect my changes (N/A —
no user-facing docs; rationale documented inline)
- [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
🤖 Generated with [Claude Code](https://claude.com/claude-code)
## Thinking Path
> - Paperclip runs Codex as a sandboxed agent with its own CODEX_HOME;
at the start of each run Paperclip merges the sandbox's auth.json with
the host's so the agent can authenticate to the Codex API
> - The merge uses `codex-auth-merge-decision.cjs`, a small CJS script
that reads two `auth.json` snapshots and exits `10` (use source copy) or
`20` (keep destination copy)
> - Until now the script only handled the **inbound** direction (host →
sandbox) with hard-coded sandbox/host semantics
> - Sandbox-to-host copy-back (**outbound**) requires the same freshness
guard but with swapped roles — and exposing both directions via a flag
requires hard-coding knowledge about which role is "sandbox" and which
is "host" into the predicate itself
> - A cleaner design: the predicate is **direction-agnostic** — it
accepts a `source` path and a `destination` path; the caller decides
direction by which argument it passes first. The script answers one
question: "should destination be replaced by source?" Exit `10` = yes
(use source); exit `20` = no (keep destination)
> - This PR refactors `codex-auth-merge-decision.cjs` to this
source/destination model, removes the `--direction` flag, and extends
coverage with a 15-row source/destination predicate matrix plus updated
extract integration tests
## Linked Issues or Issue Description
No public GitHub issue exists. The underlying problem:
**Context** — Paperclip merges `auth.json` files between sandbox and
host to keep Codex credentials in sync across agent runs. The existing
decision predicate (`codex-auth-merge-decision.cjs`) handled the inbound
direction (host→sandbox) with hard-coded sandbox/host naming. A
copy-back (outbound) direction was needed, but rather than add a
`--direction` flag that hard-codes role knowledge into the predicate,
the design was simplified: the predicate is now **direction-agnostic**,
taking `source` and `destination` paths and answering "replace
destination with source?" The freshness guard is unified: exit `10` only
when source and destination share the same `account_id`, same
subscription-kind, both carry a parseable `last_refresh`, and
`source.last_refresh > destination.last_refresh` (strict). Ties and null
freshness keep the destination copy so a spent single-use refresh token
(openai/codex#15502) is never written back over a valid one.
Related PRs:
- Refs #9785 — Phase 2: relocated the decision predicate from
`adapter-utils` into `codex-local`; this PR builds on that merged change
- Refs #9621 — community PR targeting Codex sandbox auth sync-back via a
different approach; operates on different files (runtime layer vs.
decision script), non-conflicting
## What Changed
-
`packages/adapters/codex-local/src/server/codex-auth-merge-decision.cjs`
— refactored from direction-parameterized (sandbox/host with
`--direction`) to **direction-agnostic source/destination** semantics.
Args: first = source path, second = destination path. Exit `10` (use
source) only when: same `account_id`, same subscription-kind, both carry
a parseable `last_refresh`, and `source.last_refresh >
destination.last_refresh` (strict). Every other case exits `20` (keep
destination). Unknown/extra flags are still rejected with exit `1`. No
token bytes are emitted to stdout/stderr.
- `packages/adapters/codex-local/src/server/codex-auth-merge.test.ts` —
rewritten to source/destination API: 15-row predicate matrix
(strictly-newer→`10`;
tie/older/identity-mismatch/null-refresh/apikey/kind-mismatch/unusable→`20`)
plus token-bytes non-emission guard. The inbound extract integration
suite updated to match.
- Runtime callers (`codex-auth-merge-extract.sh`) already invoke `node
decision <sandbox> <host>` = `(source, destination)` and branch on exit
`10` → keep sandbox, so **no caller changes are needed**.
**Behavioral note:** unifying the two directions under a single "use
source only when strictly fresher" rule means a tie (exact equal
`last_refresh`) now keeps the **destination** for both the inbound and
outbound frames. For inbound restore, this means an exact timestamp tie
keeps the host copy. On a tie, both credentials are the same generation
— preferring the canonical destination (host) copy is more conservative.
This is an intentional consequence of the unified guard.
## Verification
```sh
# From the codex-local package root:
pnpm vitest run src/server/codex-auth-merge.test.ts
# Expected: all predicate rows green (15-row source/destination matrix + token-leak guard)
# Plus: codex-auth-merge-extract integration rows green
```
The test suite drives the `.cjs` as a child process over temporary
`auth.json` fixtures:
- Source/destination predicate matrix (15 rows): strictly-newer→`10`;
tie→`20`; older→`20`; identity-mismatch→`20`; source last_refresh
unparseable→`20`; destination last_refresh unparseable→`20`; apikey
either side→`20`; kind-mismatch→`20`; unusable either side→`20`
- Token-bytes non-emission guard: asserts no auth token bytes appear in
`stdout`/`stderr`
- Inbound extract integration: `codex-auth-merge-extract.sh` end-to-end
rows including tie/identity-mismatch/sandbox-newer cases
## Risks
- **Tie behavior change**: exact `last_refresh` ties now always keep the
destination copy (was: keep-sandbox for inbound). On a tie both
credentials are the same generation; keeping the canonical destination
(host) copy is the more conservative choice. Covered by test assertions.
- **Token non-emission**: `stdout`/`stderr` of the `.cjs` must not echo
credential bytes; covered by the token-bytes non-emission guard test.
- **Unknown/extra flags**: handled with `process.exit(1)` rather than a
silent default, so misconfigured callers surface errors immediately.
- **Caller compatibility**: the only runtime caller,
`codex-auth-merge-extract.sh`, already calls `node decision <sandbox>
<host>` with no flags — compatible with the new source/destination API
with no changes required.
- **Low risk overall**: the script is exercised end-to-end via the
extract integration tests; all paths covered.
## Model Used
Claude claude-sonnet-4-6 (Anthropic, `claude-sonnet-4-6`), 200 k context
window, tool use and agentic code execution enabled. Automated Paperclip
multi-agent system.
## 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: Harold Kim <harold@paperclip.ing>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is an open source platform for managing AI agent teams,
with adapter packages providing concrete runtime environments (e.g.
`codex-local` runs a local OpenAI Codex sandbox)
> - `adapter-utils` is the shared utilities package — it should hold
only generic, adapter-agnostic primitives (sandbox lifecycle helpers,
merge logic, type definitions) usable by every adapter
> - Three files lived in `adapter-utils/src/` that are entirely
Codex-specific: `codex-auth-merge-extract.sh` (shell script that
extracts auth tokens), `codex-auth-merge-decision.cjs` (CJS decision
helper), and `codex-auth-merge-scripts.ts` (TypeScript factory that
wires them into the inbound provision seam added in PR #9778)
> - Their presence in a "generic" package violates the adapter isolation
principle and requires a cross-package build step to copy `.sh`/`.cjs`
files into `codex-local/dist/server/` at build time
> - This PR completes Phase 2 of the inbound-seam refactor: move all
three files to `packages/adapters/codex-local/src/server/`, rebase the
TypeScript imports, update `execute.ts` to import locally, move the
Codex-auth tests into a new `codex-local` test file, and fix packaging
so `codex-local` copies its own scripts to `dist/server/`
> - Script bytes are identical after the move; no change to inbound
auth-merge behavior or to which bytes cross the sandbox boundary
> - The benefit is a clean ownership boundary: `adapter-utils` retains
only generic runtime code, and the structural "Codex-free core" test in
the adapter-utils suite validates this invariant going forward
## Linked Issues or Issue Description
No pre-existing public GitHub issue. Describing the underlying problem
inline:
**Problem:** `packages/adapter-utils/src/` contains three files
(`codex-auth-merge-extract.sh`, `codex-auth-merge-decision.cjs`,
`codex-auth-merge-scripts.ts`) consumed exclusively by the `codex-local`
adapter. Their presence in a generic utilities package violates adapter
isolation and requires a cross-package build step (copy `.sh`/`.cjs`
into `codex-local/dist/server/`). The inbound provision seam landed in
PR #9778 routed these through `adapter-utils`; this PR finishes the
relocation.
**Related:** Refs #9778 (Phase 1 — generic asset-lifecycle-seam, now
merged).
## What Changed
- Moved `codex-auth-merge-extract.sh`, `codex-auth-merge-decision.cjs`,
and `codex-auth-merge-scripts.ts` from `packages/adapter-utils/src/` →
`packages/adapters/codex-local/src/server/` (script bytes are unchanged;
`codex-auth-merge-scripts.ts` imports rebased to `adapter-utils`
subpaths for `shellQuote` and `SandboxManagedRuntimeAssetProvision`)
- `packages/adapters/codex-local/src/server/execute.ts`: updated import
of `buildCodexAuthInboundProvision` from `adapter-utils` → local
`./codex-auth-merge-scripts`
- New
`packages/adapters/codex-local/src/server/codex-auth-merge.test.ts` — 5
Codex-auth test rows extracted from
`adapter-utils/src/workspace-restore-merge.test.ts`; the generic test
file stays and no longer references Codex
- `packages/adapters/codex-local/package.json`: build now copies
`.sh`/`.cjs` from `src/server/` into `dist/server/` directly
- `packages/adapter-utils/package.json`: removed the now-obsolete
cross-package copy step for those files
## Verification
```sh
# Type-check both affected packages
pnpm --filter @paperclipai/adapter-utils --filter @paperclipai/adapter-codex-local typecheck
# Codex auth-merge suite (moved tests — 5/5 pass)
pnpm --filter @paperclipai/adapter-codex-local test -- --reporter=verbose codex-auth-merge
# Full adapter-utils suite including the structural "core free of Codex literals" test (243 passed, 4 pre-existing skips)
pnpm --filter @paperclipai/adapter-utils test
# Confirm build copies scripts to dist/server
pnpm --filter @paperclipai/adapter-codex-local build
ls packages/adapters/codex-local/dist/server/*.sh packages/adapters/codex-local/dist/server/*.cjs
```
All of the above were run locally and passed before this PR was opened.
## Risks
Low risk — pure file relocation:
- No change to `.sh` or `.cjs` script bytes; the same content reaches
the sandbox boundary as before
- No behavioral change to inbound auth-merge logic from the sandbox's
perspective
- `adapter-utils`' structural "core free of Codex literals" test now
validates that the relocation is complete and will catch any regression
- Only `codex-local` consumed these files from `adapter-utils`; no other
package in the monorepo imported them from there
## Model Used
Claude Sonnet 4.6 (`claude-sonnet-4-6`) — Anthropic, 200k context
window, tool use, agentic reasoning. Used to author the refactor.
`Co-authored-by: Paperclip <noreply@paperclip.ing>` trailer present on
all commits.
## 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: Harold Kim <harold@paperclip.ing>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip's sandbox managed runtime is responsible for provisioning
the agent's execution environment — it extracts a home directory asset
into the sandbox before the adapter runs.
> - The sandbox runtime core was directly branching on the adapter key
(`codex`) to decide which merge scripts to stage and which merge-extract
command to run, coupling generic infrastructure to a specific adapter's
credential-merge protocol.
> - This makes it harder to add, remove, or modify per-adapter asset
provisioning without touching the runtime core; it also prevents other
adapters from contributing staged files or a custom extract command at
all.
> - The fix is to move the adapter-specific knowledge into the adapter
itself: the asset descriptor gains optional `provision` (stageFiles +
extractCommand) and `restore` contribution fields that any adapter can
populate, and the runtime core consumes them generically.
> - This pull request introduces those contribution fields, wires the
Codex adapter's inbound credential-merge as a `provision` contribution,
and removes the adapter-specific branching from the runtime core.
> - The benefit is a clean seam: the runtime core is now
adapter-agnostic for asset provisioning, the inbound behavior is
unchanged (same merge matrix, same scripts), and other adapters can
attach custom staged files or extract commands without modifying shared
infrastructure.
## Linked Issues or Issue Description
No pre-existing public GitHub issue. Describing the problem inline per
the feature template:
**Problem or motivation**
The sandbox managed-runtime asset provisioning in
`sandbox-managed-runtime.ts` branched directly on the adapter key
(`codex`) to decide which merge scripts to stage and which shell command
to use during asset extraction. This tight coupling prevents other
adapters from customizing their provisioning without modifying the
runtime core, and it means the runtime core must import and know about
adapter-specific merge scripts.
**Proposed solution**
Add an optional `provision` contribution (array of `stageFiles` entries
+ an `extractCommand` string) and an optional `restore` contribution to
the asset descriptor returned by adapters. The runtime core now consumes
these generically — if a `provision` contribution is present, it stages
those files and uses the supplied command; otherwise it falls back to
the default `tar -xf` extraction. The Codex adapter populates the
`provision` contribution where it previously depended on core branching.
**Alternatives considered**
Keeping the adapter-specific logic in the core as a documented
exception; rejected because it makes the seam inextensible.
**Roadmap alignment**
Decoupling — removes a latent coupling between the runtime core and a
specific adapter.
## What Changed
- Added `provision` contribution field (`stageFiles: Array<{src, dest}>`
+ `extractCommand: string`) to the `SandboxManagedRuntimeAsset`
descriptor type in `adapter-utils`.
- Added `restore` contribution field (hook for post-restore logic,
populated in a later phase) to the descriptor.
- Removed adapter-key branching (`if adapterKey === 'codex'`) from the
runtime core in `sandbox-managed-runtime.ts`; the core now reads
`provision.stageFiles` and `provision.extractCommand` generically.
- Extracted Codex-specific merge-script paths and the merge-extract
command into `codex-auth-merge-scripts.ts` in `adapter-utils`; the Codex
adapter's `execute.ts` now attaches them as a `provision` contribution
when it builds its managed-home asset descriptor.
- Updated `execution-target.ts` to pass the extended asset type through
to the adapter call site so the new fields are load-bearing end-to-end.
- Added seam-proving unit tests in `sandbox-managed-runtime.test.ts`:
contribution-less asset uses the default path; a non-adapter asset
round-trips the generic provision+restore seam; a structural assertion
verifies the runtime core carries no Codex-specific string literals.
- Added one test in `workspace-restore-merge.test.ts` confirming the
inbound merge matrix is unaffected.
## Verification
```bash
# Unit tests (20 pass):
npx vitest run packages/adapter-utils/src/sandbox-managed-runtime.test.ts packages/adapter-utils/src/workspace-restore-merge.test.ts
# Type-check both affected packages:
cd packages/adapter-utils && npx tsc --noEmit
cd packages/adapters/codex-local && npx tsc --noEmit
# Structural: runtime core carries no adapter string literals
grep -n 'codex\|auth\.json' packages/adapter-utils/src/sandbox-managed-runtime.ts
# Expected: zero matches
```
## Risks
**Low risk.** This is a behavior-preserving refactor: the inbound
provisioning output (which files get staged, which command runs) is
identical to before, now driven by the adapter-supplied contribution
instead of core branching. The existing inbound merge matrix tests are
the regression guard. No change to which bytes cross the sandbox
boundary. The SSH transport is untouched.
## Model Used
- **Provider:** Anthropic
- **Model:** Claude Sonnet 4.6 (`claude-sonnet-4-6`)
- **Context window:** 200 K tokens
- **Capabilities used:** tool use (file read/edit, bash execution,
Paperclip API), extended reasoning over multi-file TypeScript refactor
- **Mode:** agentic (Paperclip ACPX platform)
## 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>
Co-authored-by: Harold Kim <harold@paperclip.ing>
## Thinking Path
> - Paperclip is the open source control plane people use to manage
AI-agent companies and supervise governed work.
> - Decisions capture high-value operator judgment, and the
decision-training foundation merged in #9702 freezes that evidence for
later evaluation and learning.
> - Operators still need the UI from the closed stacked PR #9718 to
intentionally capture examples and inspect the resulting dataset.
> - GitHub automatically closed#9718 when its stacked base branch was
deleted after #9702 merged, leaving the server foundation on `master`
without the corresponding UI.
> - This pull request restores the final UI and its still-required
supporting API fields directly on current `master`, while excluding the
obsolete migration and duplicated server-foundation diffs.
> - The benefit is a reviewable replacement PR that preserves the
completed decision-training workflow without replaying stale stack
history.
## Linked Issues or Issue Description
- Refs #9718
- Refs #9702
## What Changed
- Restored the top-level `/training` library and record inspector with
search, filters, JSONL export, notes editing, and evidence tabs.
- Restored the Decisions-row training affordance and capture drawer,
including preview, provenance, deletion, cache refresh, and approval
consistency behavior.
- Restored the shared types and focused server support needed by the UI
without reintroducing decision-training migrations or the already-merged
server foundation.
- Restored focused UI and attention-service tests from the final #9718
state.
- Credit to the authors and reviewers of #9718; this recovery
transplants their final reviewed delta after the stacked base deletion.
## Verification
- `pnpm exec vitest run ui/src/pages/Training.test.tsx
ui/src/components/DecisionTrainingDrawer.test.tsx
ui/src/components/AttentionQueueRow.test.tsx
server/src/__tests__/attention-service.test.ts
server/src/__tests__/decision-training.test.ts` — 5 files, 48 tests
passed.
- `pnpm --filter @paperclipai/shared typecheck` — passed.
- `pnpm --filter @paperclipai/server typecheck` — passed.
- `pnpm --filter @paperclipai/ui typecheck` — passed.
- `pnpm exec vitest run server/src/__tests__/openapi-routes.test.ts` — 3
tests passed; confirms exact route/OpenAPI parity for the restored
preview endpoint.
- `pnpm check:token-gates` — the restored files are clean; the
repository-wide command currently reports five pre-existing false
positives where comments reference GitHub issue `#9627` as if it were a
color literal.
## Risks
- Low migration risk: this PR contains no database migrations and is
based directly on current `master`.
- The main behavioral risk is cache invalidation across Decisions and
Training views; focused tests cover capture, update, deletion, row
state, and approval refresh behavior.
- The token-gate baseline remains red on unrelated `#9627` comment
references; this PR does not modify those files.
> 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, reasoning with repository/tool access and
code execution. Context window not exposed by the 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 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 Inbox lists issues as rows; unread rows show a small blue
mark-read dot at the leading edge
> - The dot was rendered as an `order-first` flex column that existed
only on unread rows, so it changed each row's leading layout depending
on read state
> - On parent (chevron) rows the dot stacked on top of the collapse
chevron and pushed the whole row's status/title to the right, so unread
rows sat indented relative to their read siblings
> - This PR gives every inbox row a reserved leftmost dot slot so the
dot never changes horizontal alignment
> - The benefit is a stable, non-shifting inbox: read and unread rows
line up at the same x, and the dot no longer crowds the collapse caret
## Linked Issues or Issue Description
No public issue exists. Describing as a bug report:
- **What happened:** In the Inbox, an unread issue row's blue mark-read
dot pushed the row's status/title to the right, so unread rows were
indented relative to read siblings. On parent (collapsible) rows the dot
also overlapped the collapse chevron.
- **Expected behavior:** The unread dot should not change a row's
horizontal alignment. Read and unread rows should line up identically,
and the dot should sit clear of the collapse caret.
- **Steps to reproduce:** Open the Inbox with a mix of read/unread
issues, including a parent issue with children. Observe unread rows
indented vs. read rows, and the dot crowding the chevron on parent rows.
- **Deployment mode:** UI-only (frontend).
Related PR (same inbox area, in flight):
https://github.com/paperclipai/paperclip/pull/9685
## What Changed
- `ui/src/components/IssueRow.tsx`: Add a reserved leftmost dot slot
(`w-4`, desktop) that is present on **all** inbox rows — read and unread
alike — so the mark-read dot renders far-left ahead of any leading
control without shifting content. The dot button markup is now shared
between the desktop reserved slot and the mobile in-flow rendering. The
row is `relative` for slot positioning.
- `ui/src/pages/Inbox.tsx`: Always reserve the leading spacer for
non-chevron rows (removed the special-case that dropped it for unread
rows), since the dot now has its own dedicated slot to the left of the
spacer. This removes the double-indent.
- Mobile keeps the dot in flow as the leading item (no reserved desktop
gutter there).
- Tests: `IssueRow.test.tsx` and `Inbox.test.tsx` updated to assert the
reserved dot slot exists on read and unread rows and that alignment no
longer differs by read state.
## Verification
- `cd ui && NODE_ENV=development npx vitest run
src/components/IssueRow.test.tsx src/pages/Inbox.test.tsx` → 28 tests
pass.
- `cd ui && npm run typecheck` → clean.
- Note: running the full UI suite with `NODE_ENV=production` surfaces
pre-existing `act is not a function` failures unrelated to this change
(React's production bundle strips `act`); they pass under a normal test
env.
## Risks
Low risk. UI-only, scoped to inbox row layout. The reserved slot is
`hidden ... sm:inline-flex`, so mobile layout is unchanged (dot stays in
flow). No API, schema, or telemetry changes.
## Model Used
Claude Opus 4.8 (claude-opus-4-8), 1M context window, 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 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
- [ ] I have updated relevant documentation to reflect my changes (no
docs affected)
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green (pending CI)
- [ ] 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>
## Thinking Path
> - Paperclip is the open source control plane people use to manage
AI-agent companies and their work
> - Human approvals, issue interactions, and execution decisions already
capture high-value decision moments
> - Those moments are currently transient and cannot be reused as stable
evaluation or training examples
> - Reusable examples need a server-owned, immutable snapshot so later
comments or runs cannot leak into the recorded state
> - Human notes need to remain editable and auditable without changing
the captured state
> - This pull request adds the database model, snapshot capture service,
API, export format, and attention-feed enrichment for decision training
> - The benefit is a durable, inspectable foundation for evaluating
whether agents can reproduce good human decisions from only the context
available at decision time
## Linked Issues or Issue Description
### Subsystem affected
Cross-cutting (`server/`, `packages/db`, and `packages/shared`).
### Problem or motivation
Paperclip has no durable dataset for converting human decisions into
evaluation-ready examples. Teams need to capture pending or resolved
decisions with the exact issue context, comments, runs, and repository
evidence available at a cutoff, while preventing future context from
leaking into the example.
### Proposed solution
Store immutable, schema-versioned snapshots anchored to durable
interaction, approval, or execution-decision records; keep notes
separately editable with history; expose human-only CRUD, list, and
JSONL export APIs.
### Alternatives considered
Client-generated snapshots were rejected because they duplicate cutoff
logic and cannot reliably enforce no-leakage boundaries. Automatic
outcome backfill was deferred so captured examples remain faithful to
what was known at capture time.
### Roadmap alignment
Supports the roadmap direction of turning completed work and decision
patterns into reusable organizational knowledge.
### Additional context
The implementation records explicit commit-resolution confidence
(`exact`, `nearest_run`, `workspace`, or `none`) so downstream
evaluation can distinguish evidence quality.
## What Changed
- Added the `decision_training_examples` schema and idempotent migration
with company, issue, and source/author indexes.
- Added shared types for decision-training records, notes history, and
versioned snapshots.
- Added a single server-side snapshot capture path with inclusive
comment cutoffs, pre-cutoff run capture, durable decision payloads, and
explicit commit-resolution confidence.
- Added create, list, detail, notes-only update, delete, and JSONL
export routes with human-only write authorization and activity logging
that skips no-op note submissions.
- Added per-user `trainingExampleId` enrichment to attention items.
- Added focused embedded-Postgres tests for cutoff boundaries,
post-cutoff leakage, immutable snapshots, human-only writes, duplicate
prevention, notes history, attention enrichment, and export shape.
- Updated UI test and Storybook attention-item factories for the new
required `trainingExampleId` contract.
## Verification
- `pnpm exec vitest run server/src/__tests__/decision-training.test.ts`
— 10 tests passed.
- `pnpm --filter @paperclipai/db typecheck` — passed, including
migration numbering and safety checks.
- `pnpm --filter @paperclipai/shared typecheck` — passed.
- `pnpm --filter @paperclipai/server typecheck` — passed.
- `pnpm --filter @paperclipai/ui typecheck` — passed.
## Risks
- The migration adds a new table and indexes only; it does not rewrite
existing rows or install resolve-time hooks.
- Snapshot JSON can grow with long comment threads and run histories; v1
intentionally favors complete, inspectable examples over aggressive
truncation.
- Commit SHA resolution is evidence-based and records `exact`,
`nearest_run`, or `none` so downstream consumers can account for
confidence.
- The API is additive, but future UI work must continue to treat the
snapshot as immutable and use notes-only updates.
> 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 using `gpt-5.3-codex`, with repository tool use, terminal
execution, and code-editing capabilities; context-window size is not
exposed by the 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 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 is the open source control plane people use to organize,
govern, and understand AI-agent work
> - Operators need concise, current status views across projects and
execution workspaces without manually reading every issue and run
> - Paperclip already has auditable issues, documents, built-in agents,
routines, and live run events, but no first-class summary-slot workflow
connecting those systems
> - A built-in Summarizer can generate status prose through ordinary
governed tasks while summary slots provide stable, revisioned
destinations for that output
> - The UI needs to show current summaries, generation progress,
failures, revisions, and streaming draft status in the places operators
already work
> - This pull request adds the end-to-end summary-slot data, API, agent,
orchestration, and UI surfaces behind an experimental setting
> - The benefit is decision-oriented status context that remains
company-scoped, auditable, retryable, and inexpensive by default
## Linked Issues or Issue Description
No public GitHub issue exists for this feature.
**Problem**
Operators currently have to reconstruct project and workspace status by
reading many issues, runs, and comments. This makes it hard to identify
decisions, review queues, recent work, and the next event worth
watching.
**Proposed capability**
Add an experimental summary system with revisioned summary slots for
projects and workspaces, a paused-by-default built-in Summarizer agent,
governed generation tasks, live draft status, and reusable UI cards.
**Expected behavior**
- Summary data remains company-scoped and revisions remain auditable.
- Generation runs through normal issue/agent orchestration and
deduplicates active requests.
- Only the linked built-in Summarizer generation task can author a slot
revision.
- Operators can generate, retry, inspect revisions, and follow draft
progress from project and workspace views.
- The feature remains opt-in and background generation remains paused by
default.
## What Changed
- Added summary-slot schema, idempotent migrations, shared contracts,
validators, API paths, and service tests.
- Added company-scoped summary-slot routes for reading revisions,
requesting generation, and guarded Summarizer writes with activity
logging.
- Added terminal generation finalization, failure reasons, assignment
wakeups, and orchestration integration.
- Added the paused-by-default built-in Summarizer bundle, low-cost
runtime defaults, status-summarization skill, and stale-summary routine.
- Added summary cards, revision selection, retry/configuration states,
live draft streaming, transcript chunk handling, and project/workspace
integrations.
- Updated Claude local parsing for streamed status output and expanded
server, adapter, shared, database, catalog, and UI coverage.
## Verification
- `pnpm -r typecheck`
- `pnpm exec vitest run packages/db/src/summary-slots-schema.test.ts
packages/shared/src/summary-slot.test.ts
server/src/__tests__/summary-slot-routes.test.ts
server/src/__tests__/summary-slots.test.ts
server/src/__tests__/built-in-agents.test.ts
ui/src/components/SummarySlotCard.test.tsx
ui/src/components/SummarySlotCard.status.test.tsx
ui/src/components/useSummaryDraftStream.test.tsx
ui/src/lib/summary-draft-stream.test.ts
ui/src/lib/run-log-chunks.test.ts
ui/src/context/LiveUpdatesProvider.hook.test.tsx` — 113 tests passed
- `pnpm test:run` — server and UI suites passed; one CLI AWS doctor test
was affected by inherited `AWS_ACCESS_KEY_ID`/`AWS_SECRET_ACCESS_KEY`,
and passed when those host credentials were removed
- `pnpm exec vitest run cli/src/__tests__/secrets.test.ts` with
inherited AWS credential variables removed — 8 tests passed
- `pnpm build`
- `pnpm check:token-gates` currently reports nine `#9627` comment
references introduced by current `master`; none are in this PR diff
## Risks
- Database risk is limited by incrementally ordered, idempotent
migrations and migration safety checks.
- Summary generation creates normal issues/runs, so misconfiguration can
produce failed slots; the UI exposes retryable failure reasons and agent
configuration entry points.
- Streaming draft parsing depends on the documented `STATUS:` protocol;
final persisted revisions remain the source of truth.
- The feature is experimental, opt-in, and its built-in routine is
paused with no background token spend by default.
> 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 GPT-5.3 Codex with reasoning, repository tool use, code
execution, GitHub CLI, and Paperclip control-plane integration. Earlier
branch commits also record Claude model co-authorship where applicable.
## 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>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - The Inbox lets operators scan task state and distinguish unread
activity at a glance
> - Read and unread rows should keep the same status-and-title alignment
so the list remains easy to scan
> - The mark-read dot previously participated in flex layout, which
added an extra leading column and visibly indented unread rows
> - Parent rows also needed special handling because their collapse
chevron occupies the normal leading gutter
> - This pull request moves the unread dot out of desktop flex flow,
preserves a consistent leading spacer, and positions parent-row dots
after the chevron
> - The benefit is consistent alignment for read, unread, leaf, nested,
and collapsible parent rows without hiding tree controls
## Linked Issues or Issue Description
### What happened?
In the Inbox, an unread row rendered its status icon and title farther
right than an equivalent read row because the mark-read dot occupied its
own flex column. On a collapsible parent row, the dot also competed with
the leading chevron.
### Expected behavior
Read and unread Inbox rows keep identical status/title alignment while
preserving the unread affordance and any tree-expansion controls.
### Steps to reproduce
1. Run Paperclip from `master` and open the Inbox at desktop width.
2. Compare otherwise-equivalent read and unread leaf rows.
3. Expand a task tree containing an unread depth-zero parent.
4. Observe the unread row indentation and the dot competing with the
parent chevron.
### Paperclip version or commit
Reproduced against the pre-fix `master` parent of this PR.
- **Deployment mode:** Local dev (`pnpm dev`).
- **Installation method:** Built from source.
- **Agent adapters involved:** Not adapter-specific; this is a core
Inbox UI bug.
- **Database mode:** Not database-related.
- **Access context:** Board (human operator).
- **Additional context:** No logs or configuration are involved; the
regression is visual layout behavior covered by focused component/page
tests.
## What Changed
- Render the desktop unread dot as an absolute overlay so it does not
consume row width; retain the existing in-flow behavior on mobile.
- Add an `unreadDotPlacement` option so depth-zero collapsible parents
place the dot after the leading chevron.
- Reserve the same leading spacer for read and unread non-chevron Inbox
rows.
- Expand `IssueRow` and Inbox tests to cover leaf alignment, nested
rows, parent chevrons, fading state, and mobile behavior.
## Verification
- `pnpm exec vitest run ui/src/components/IssueRow.test.tsx
ui/src/pages/Inbox.test.tsx` — 2 files, 30 tests passed.
- `pnpm check:token-gates` — all three token gates clean across 630
scanned files.
- Browser QA (desktop 1280px and mobile 375px) — PASS: read/unread
desktop content measured at identical x positions; nested guides, parent
chevron hit target, fading state, and mobile mark-read hit targets
verified.
## Risks
- Low risk: the change is isolated to Inbox row presentation and has
focused regression coverage.
- The main visual risk is breakpoint-specific placement of the mark-read
dot; tests explicitly cover desktop absolute positioning and mobile
in-flow positioning.
> 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-5.6-sol`, high reasoning effort, with
repository/tool execution. The runtime did not expose a context-window
size.
- Original implementation commit also records assistance from Claude
Opus 4.8.
## 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 is the open source control plane people use to coordinate
AI-agent work.
> - Execution workspaces provide the local service loop where operators
start, stop, restart, inspect, and open workspace services.
> - The existing header exposed those actions through separate controls
whose position and labeling changed across runtime states.
> - That movement made the most common development actions harder to
scan and easier to misclick, especially with multiple services or long
URLs.
> - The design therefore uses one fixed-geometry, state-aware control
bar and keeps service-specific detail behind a compact disclosure.
> - This pull request adds that control surface, maps existing runtime
data and every pending mutation into it, and integrates it into the
execution-workspace header without changing server contracts.
> - The benefit is a calmer, predictable service-control loop across
stopped, transitional, running, unhealthy, failed, multi-service, and
narrow-width states.
## Linked Issues or Issue Description
### Subsystem affected
`ui/ — React + Vite board UI`
### Problem or motivation
Execution-workspace service actions move and change shape as runtime
state changes, while URLs and multi-service status compete for header
space. During bulk actions, operators also need every targeted service
to show its transitional state immediately.
### Proposed solution
Use one fixed-geometry, state-aware service control bar in the workspace
header. Map existing runtime records into a stable status, URL, and
actions model, and track each in-flight bulk request independently until
it settles.
### Alternatives considered
Keeping the separate quick-control buttons was rejected because their
geometry changes by state. Showing every service inline was rejected
because it makes the header too wide; per-service detail remains in a
compact disclosure and the Services tab.
### Roadmap alignment
Reviewed `ROADMAP.md`; this focused execution-workspace UI improvement
does not duplicate a listed roadmap initiative.
### Additional context
The published design and state viewer is available at
https://pages.paperclip.ing/pap-14233-workspace-service-controls/.
## What Changed
- Added `WorkspaceServiceControlBar`, a fixed-geometry responsive
control for single- and multi-service runtime states.
- Added 15 Storybook states covering running, stopped, transitions,
unhealthy, failed, disabled, long-URL, mobile, and multi-service
behavior.
- Replaced `WorkspaceRuntimeQuickControls` in the execution-workspace
header with adapters that map live services and all pending requests
into the new control model.
- Added focused unit coverage for service-entry construction, bulk
pending overlays, request resolution, clipboard feedback, and header
integration.
## Verification
- `cd ui && NODE_ENV=development pnpm vitest run
src/components/WorkspaceServiceControlBar.test.tsx
src/components/WorkspaceRuntimeControls.test.tsx
src/pages/ExecutionWorkspaceDetail.test.tsx` — 29 tests passed.
- `NODE_ENV=development pnpm --dir ui typecheck` — passed.
- `pnpm check:token-gates` — all token gates clean.
- Reviewed the Storybook captures for all primary states.
## Risks
- Low-to-moderate UI risk: service controls depend on adapter mapping
from existing runtime records; focused tests cover single-service and
bulk-action mapping and integration paths.
- Multi-service bulk actions intentionally apply to all eligible
services, while per-service actions remain in the disclosure.
> 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, exact model ID `gpt-5.3-codex`; runtime-managed context
window; coding/reasoning mode with repository, terminal, Git, GitHub
CLI, and test execution tools.
## 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: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the control plane people use to manage AI-agent
companies and their work
> - The activity log is the generic audit spine for mutations across the
control plane
> - Activity rows identify agents and runs, but they do not persist the
responsible human upstream
> - Relying only on run joins loses attribution after run pruning and
misses agent API-key actions outside a run
> - This pull request resolves responsible-user attribution when each
activity row is written and stores it directly
> - The benefit is durable, queryable agent audit feeds without
rewriting historical provenance
## Linked Issues or Issue Description
### Problem or motivation
Agent activity records do not persist the responsible user, so
attribution can disappear when runs are pruned and no-run API-key
mutations cannot be attributed correctly.
### Proposed solution
Resolve attribution for each new activity row from the run, related
issue, active agent API key, or company default, in that order, and
persist the result directly.
### Alternatives considered
Read-time joins alone were rejected because pruned runs lose durable
attribution and out-of-run agent-key actions have no run to join.
Historical backfill was rejected because it would invent provenance.
### Roadmap alignment
This strengthens the durable audit-trail direction described in
`ROADMAP.md` without adding a new product surface.
## What Changed
- Added nullable `activity_log.responsible_user_id` plus
company/agent/time and company/responsible-user/time indexes.
- Added an idempotent forward-only migration with no historical
backfill.
- Added centralized write-time resolution: heartbeat run → issue
attribution → active agent API key → company default.
- Propagated authenticated API-key IDs through existing request-backed
`logActivity()` calls.
- Added unit coverage for every fallback and an embedded-Postgres
assertion for the no-run API-key stamping path.
## Verification
- `pnpm --filter @paperclipai/db typecheck`
- `pnpm --filter @paperclipai/server typecheck`
- `pnpm --filter @paperclipai/server exec vitest run
src/__tests__/activity-log-responsible-user.test.ts
src/__tests__/authz-company-access.test.ts`
- Initial focused verification: 25 tests passed; migration safety
passed.
- Follow-up regression verification: 54 focused
attribution/company-skill/environment/issue-tree/tool-gateway tests
passed; server typecheck passed.
- GitHub: full build, typecheck, server shards, serialized suites, e2e,
security, and policy checks passed.
## Risks
- Adding two indexes to an existing large table can hold a write lock
while the transactional migration runs. The migration safety
suppressions document why `CONCURRENTLY` is unavailable under the
current Drizzle migration runner.
- Historical rows remain nullable by design; this avoids inventing
provenance and keeps the migration forward-only.
- API-key attribution requires request-backed activity call sites to
pass the authenticated key ID; this PR mechanically updates the existing
actor-based activity calls and covers the no-run path with integration
testing.
> 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-5.4`; context-window size was not exposed by the
runtime. Medium reasoning with repository editing, terminal execution,
and test execution capabilities was used.
## 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 control plane people use to coordinate
AI-agent companies
> - Agents, routines, productivity review, and recovery services can all
create or re-trigger work
> - Repeated heartbeats or catch-up cycles can produce duplicate tasks
or repeat recovery actions before prior work is visible
> - The base duplicate-create and routine catch-up coalescing work has
now landed on `master` via related PRs while this PR was being prepared
> - This pull request carries the remaining hardening: bounded
idempotency retention, recovery cooldown/throttle fixes,
productivity-review query batching and ordering fixes, and regression
coverage
> - The benefit is fewer duplicate tasks, safer retries, and enough
provenance to diagnose any future recurrence
## Linked Issues or Issue Description
Agents can retry issue creation after ambiguous responses or
independently recreate the same child title, while recovery and
short-interval routine catch-up paths can repeat before prior work
settles. This can produce visible duplicate tasks and makes the
originating heartbeat difficult to identify.
Related work: Refs #8356 for caller-supplied issue-create idempotency
and Refs #9224 for plugin-scoped issue-create idempotency. Prior related
PR: #6936. The base issue-create deduplication and routine catch-up
coalescing pieces have since landed on `master` via #9650 and #9649;
this PR remains as the follow-up hardening stack on top of those
changes.
## What Changed
- Add 7-day retention for issue-create idempotency claims with indexed,
batch-limited cleanup so the claim table does not grow forever.
- Preserve recovery cooldown intent after terminal recovery actions are
closed, and throttle repeated source-scoped recovery work.
- Batch productivity-review source-activity checks to avoid repeated
per-source queries while keeping the no-action suppression behavior.
- Order productivity-review no-action streak windows by review creation
time, matching the window semantics even when completion timestamps are
out of order.
- Preserve generated issue IDs in route mocks used by backlog/assignment
contract tests.
- Document PR-gardening task deduplication expectations in the company
skill.
- Add focused regression tests for idempotency retention, liveness
recovery cooldowns, and productivity-review
batching/suppression/ordering behavior.
## Verification
- `pnpm --filter @paperclipai/server typecheck` — passed on latest head.
- `pnpm exec vitest run --project @paperclipai/server
server/src/__tests__/productivity-review-service.test.ts
server/src/__tests__/issue-create-deduplication-routes.test.ts
--reporter verbose` — 2 files, 23 tests passed on latest head.
- `pnpm --filter @paperclipai/adapter-utils build && pnpm exec vitest
run --project @paperclipai/adapter-utils --reporter dot` — 30 files
passed, 476 tests passed, 8 skipped.
- `pnpm build` — passed on latest head.
- `pnpm -r typecheck` — passed on the rebased head before the final
productivity-review ordering commit; the latest touched server code is
covered by the server typecheck above.
- `pnpm check:token-gates` — passed.
- `pnpm test:run` — progressed through server, UI, CLI, shared,
skills-catalog, and DB sections, then exposed an adapter-utils
compiled-test fingerprint mismatch before rebuilding adapter-utils; the
adapter-utils project passed after rebuild, and the GitHub split PR
checks passed on the pushed head.
## Risks
- Caller-supplied idempotency replay is now bounded to 7 days; reusing
an old key after retention can create new work, which matches
retry-oriented idempotency semantics.
- Recovery and productivity-review timing changes may suppress redundant
follow-up work; focused tests cover the intended boundaries.
- Advisory locking and idempotency cleanup rely on PostgreSQL-compatible
transaction semantics already used by the production data layer.
- The migration extends the private claim table indexes without
rewriting existing issue rows.
> 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 reasoning, repository editing,
shell/tool execution, and test execution. Exact model ID and
context-window size are not exposed by the 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 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>
Nicky Leach
Sam
Harshit Khemani
dependabot[bot]
le-czs
Stefano Maffeis
Harsh Kotak
馨冉
Dotta
Devin Foley
openclaw-fmag
NoSoloSoft
