A failed final flush does not erase an earlier successful periodic checkpoint. Cover interrupted continuations and replacement sandboxes for both saved and failed prior runs.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Report incomplete terminal saves without hiding the latest successful checkpoint, including same-sandbox continuation and replacement failures.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Fail before editing repository state if the remote cannot be read, and reject credential or PATH workarounds during model-driven acceptance.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Align durable journal validation with the transport bound, preserve authorization on cached storage clients, and avoid Cloud session gates for native Git callbacks.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Defer cached-file debugging and live sandbox inspection as separate features. Keep the saved-file browser only as an unshipped Storybook prototype.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Require a reviewed lock digest for staging migrator and app builds so registry drift fails before lifecycle-enabled installation.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Preserve migration SQL hashes when renumbering and keep sandbox HOME with managed GitHub shell profiles.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Wait for the final work-folder checkpoint before background reconciliation can finalize a sandbox run. Persist its resumable task identity before exposing completion, and avoid late cleanup overwriting a newer turn.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Wait for the final work-folder checkpoint before background reconciliation can finalize a sandbox run. Persist its resumable task identity before exposing completion, and avoid late cleanup overwriting a newer turn.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Keep the unverified backend closed while allowing the descriptor-confined runner path, with regression coverage through the production session entry point.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Codex filters explicit environment overrides through include_only. Retain the scoped home and standard executable path without allowing provider or host secrets into shell commands.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Codex filters explicit environment overrides through include_only. Retain the scoped home and standard executable path without allowing provider or host secrets into shell commands.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - The recent tasks sidebar helps operators return to active and
completed work.
> - Task detail refreshes update the activity timestamps used to sort
that list.
> - Older responses can undo newer activity. Concurrent updates can move
rows repeatedly.
> - This pull request preserves the newest observed activity and
debounces row moves.
> - Operators can select a stable row while task text and live state
stay current.
## Linked Issues or Issue Description
**What happened?**
Recent task rows can repeatedly swap positions when two tasks receive
updates. Older detail responses can also move a task below another task
after a newer comment promoted it.
**Expected behavior**
Rows remain stable during an update burst. The final activity order
appears after one second of quiet. Old responses never reduce the
recorded activity time.
**Steps to reproduce**
1. Enable the streamlined UI and open multiple tasks.
2. Send alternating updates to two tasks in the recent list.
3. Observe row order while those updates arrive and while older detail
data refreshes.
**Paperclip version or commit**
Base commit: 1cc45086d.
**Deployment mode**
Board UI. The behavior is independent of the server deployment mode.
Related work: #12854 and #12746 introduced the current navigation.
#12314 is an earlier experimental sidebar proposal; this fix targets the
implementation already on master.
## What Changed
- Keep stored activity timestamps at their newest observed value in both
recording paths.
- Debounce activity-driven row moves for one second. Keep additions and
removals immediate.
- Keep detail query observers in a fixed order and resolve task text by
task ID.
- Add regression coverage for alternating updates, stale data,
additions, and removals.
- Document recent task ordering in the development guide.
## Verification
- Passed: 17 focused recent task tests.
- Passed: `pnpm check:token-gates`.
- Passed: `pnpm -r typecheck` and `pnpm build`.
- Passed: Greptile 5/5 on commit
`8861961de9aaa7e9a8184826bc5ce367450f9e8a`, with no review threads or
actionable findings.
- Passed: all 31 remote checks, including all general and serialized
test shards, browser tests, release verification, and build. The opt-in
Storybook visual check was skipped by policy.
- One unrelated concurrent artifact-document test failed on the first CI
attempt. Its 11-test suite passed locally, and the single CI retry
passed without code changes.
- The duplicate full local `pnpm test:run` was started; the completed CI
suite provides the full-suite result.
- Regression tests use fake timers to check that rows remain fixed
during alternating activity, text refreshes immediately, and the final
order appears after one second of quiet.
## Risks
- Continuous activity intentionally delays row moves until updates
settle. New and removed tasks still appear immediately.
- Storage format and server contracts stay compatible. No migration is
needed.
## Model Used
OpenAI GPT-6 through Codex, with reasoning, repository inspection, code
editing, and test execution. The exact deployment model ID and context
window size are not exposed in this 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 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>
Persist the validated workspace mode and run identical executable acceptance scripts across live engines.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Persist the validated workspace mode and run identical executable acceptance scripts across live engines.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Dotta
Dotta