## Thinking Path
> - Paperclip orchestrates AI agents for zero-human companies.
> - Heartbeat runs are the control-plane record of each agent execution
window.
> - Long-running local agents can exhaust context or stop while still
holding useful next-step state.
> - Operators need that stop reason, next action, and continuation path
to be durable and visible.
> - This pull request adds run liveness metadata, continuation
summaries, and UI surfaces for issue run ledgers.
> - The benefit is that interrupted or long-running work can resume with
clearer context instead of losing the agent's last useful handoff.
## What Changed
- Added heartbeat-run liveness fields, continuation attempt tracking,
and an idempotent `0058` migration.
- Added server services and tests for run liveness, continuation
summaries, stop metadata, and activity backfill.
- Wired local and HTTP adapters to surface continuation/liveness context
through shared adapter utilities.
- Added shared constants, validators, and heartbeat types for liveness
continuation state.
- Added issue-detail UI surfaces for continuation handoffs and the run
ledger, with component tests.
- Updated agent runtime docs, heartbeat protocol docs, prompt guidance,
onboarding assets, and skills instructions to explain continuation
behavior.
- Addressed Greptile feedback by scoping document evidence by run,
excluding system continuation-summary documents from liveness evidence,
importing shared liveness types, surfacing hidden ledger run counts,
documenting bounded retry behavior, and moving run-ledger liveness
backfill off the request path.
## Verification
- `pnpm exec vitest run packages/adapter-utils/src/server-utils.test.ts
server/src/__tests__/run-continuations.test.ts
server/src/__tests__/run-liveness.test.ts
server/src/__tests__/activity-service.test.ts
server/src/__tests__/documents-service.test.ts
server/src/__tests__/issue-continuation-summary.test.ts
server/src/services/heartbeat-stop-metadata.test.ts
ui/src/components/IssueRunLedger.test.tsx
ui/src/components/IssueContinuationHandoff.test.tsx
ui/src/components/IssueDocumentsSection.test.tsx`
- `pnpm --filter @paperclipai/db build`
- `pnpm exec vitest run server/src/__tests__/activity-service.test.ts
ui/src/components/IssueRunLedger.test.tsx`
- `pnpm --filter @paperclipai/ui typecheck`
- `pnpm --filter @paperclipai/server typecheck`
- `pnpm exec vitest run server/src/__tests__/activity-service.test.ts
server/src/__tests__/run-continuations.test.ts
ui/src/components/IssueRunLedger.test.tsx`
- `pnpm exec vitest run
server/src/__tests__/heartbeat-process-recovery.test.ts -t "treats a
plan document update"`
- `pnpm exec vitest run server/src/__tests__/activity-service.test.ts
server/src/__tests__/heartbeat-process-recovery.test.ts -t "activity
service|treats a plan document update"`
- Remote PR checks on head `e53b1a1d`: `verify`, `e2e`, `policy`, and
Snyk all passed.
- Confirmed `public-gh/master` is an ancestor of this branch after
fetching `public-gh master`.
- Confirmed `pnpm-lock.yaml` is not included in the branch diff.
- Confirmed migration `0058_wealthy_starbolt.sql` is ordered after
`0057` and uses `IF NOT EXISTS` guards for repeat application.
- Greptile inline review threads are resolved.
## Risks
- Medium risk: this touches heartbeat execution, liveness recovery,
activity rendering, issue routes, shared contracts, docs, and UI.
- Migration risk is mitigated by additive columns/indexes and idempotent
guards.
- Run-ledger liveness backfill is now asynchronous, so the first ledger
response can briefly show historical missing liveness until the
background backfill completes.
- UI screenshot coverage is not included in this packaging pass;
validation is currently through focused component 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`.
## Model Used
- OpenAI Codex, GPT-5.4, local tool-use coding agent with terminal, git,
GitHub connector, GitHub CLI, and Paperclip API access.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
Screenshot note: no before/after screenshots were captured in this PR
packaging pass; the UI changes are covered by focused component tests
listed above.
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
The Codex CLI can return "no rollout found for thread id ..." when
resuming a heartbeat thread whose rollout has been garbage-collected.
Extend isCodexUnknownSessionError() to match this wording so the
existing single-retry path in execute.ts activates correctly.
Add parse.test.ts covering the new pattern, existing stale-session
wordings, parseCodexJsonl, and a negative case.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
After the mocked RPC spawn fails, getQuotaWindows() still calls
readCodexToken(). Use an empty mkdtemp directory for CODEX_HOME for the
duration of the test so we never read ~/.codex/auth.json or call WHAM.
Add a Vitest case that mocks `node:child_process.spawn` so the child
emits `error` (ENOENT) after the constructor attaches listeners.
`getQuotaWindows()` must resolve with `ok: false` instead of leaving an
unhandled `error` event on the process.
Register `packages/adapters/codex-local` in the root Vitest workspace.
Document in DEVELOPING.md that a missing `codex` binary should not take
down the API server during quota polling.
When the `codex` binary is absent from PATH, Node.js emits an `error`
event on the ChildProcess. Because `CodexRpcClient` only subscribed to
`exit` and `data` events, the `error` event was unhandled — causing
Node to throw it as an uncaught exception and crash the server.
Add an `error` handler in the constructor that rejects all pending RPC
requests and clears the queue. This makes a missing `codex` binary a
recoverable condition: `fetchCodexRpcQuota()` rejects, `getQuotaWindows()`
catches the error and returns `{ ok: false }`, and the server stays up.
The fix mirrors the existing pattern in `runChildProcess`
(packages/adapter-utils/src/server-utils.ts) which already handles
`ENOENT` the same way for the main task execution path.
The previous documentation parenthetical "(defaulting to ~/.codex/skills/)"
was misleading because Paperclip almost always sets CODEX_HOME to a
per-company managed home. Update index.ts docs, skills.ts detail string,
and execute.ts inline comment to make the runtime path unambiguous.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
The default fallback in ensureCodexSkillsInjected still referenced the
old function name. Updated to use resolveCodexSkillsDir with shared
home as fallback.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
The previous commit incorrectly used resolveSharedCodexHomeDir() (~/.codex)
but Codex runs with CODEX_HOME set to a per-company managed home under
~/.paperclip/instances/. Skills injected into ~/.codex/skills/ would not
be discoverable by Codex. Now uses effectiveCodexHome directly.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
The Codex adapter was the only one injecting skills into
<cwd>/.agents/skills/, polluting the project's git repo. All other
adapters (Gemini, Cursor, etc.) use a home-based directory. This
changes the Codex adapter to inject into ~/.codex/skills/ (resolved
via resolveSharedCodexHomeDir) to match the established pattern.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Use path.join instead of string concatenation for the auth.json
fallback path in the detail message, ensuring correct path
separators on Windows.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When adapter config has no CODEX_HOME but process.env.CODEX_HOME is
set, readCodexAuthInfo reads from the process env path. The detail
message now uses codexHomeDir() instead of hardcoded "~/.codex" so
the displayed path always matches where credentials were read from.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Show the configured CODEX_HOME path instead of hardcoded ~/.codex
when the email fallback message is displayed.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add tests for codex_native_auth_present and codex_openai_api_key_missing
code paths. Also pass adapter-configured CODEX_HOME through to
readCodexAuthInfo so the probe respects per-adapter home directories.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Loading an instructions file is normal, expected behavior — not worth
logging to stdout/stderr on every run. Warning logs for failed reads
are preserved.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
The environment test warned about OPENAI_API_KEY being unset even
when Codex was authenticated via `codex auth`. Now checks
~/.codex/auth.json before emitting the warning.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
mkdir -p the CODEX_HOME directory in codex-local adapter and the
agentHome directory in the heartbeat service before passing them to
adapters. This prevents CLI tools from erroring when their home
directory hasn't been created yet. Covers all local adapters that
set AGENT_HOME.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Persist child-process metadata for local adapter runs, keep detached runs alive when their pid still exists, queue a single automatic retry when the pid is confirmed dead, and clear detached warnings when the original run reports activity again.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Expose adapter-discovered user-installed skills with provenance metadata, share persistent skill snapshot classification across local adapters, and render unmanaged skills as a read-only section in the agent skills UI.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
- Extract all Anthropic credential/API logic into claude-local/src/server/quota.ts
- Extract all OpenAI/WHAM credential/API logic into codex-local/src/server/quota.ts
- Add optional getQuotaWindows() to ServerAdapterModule in adapter-utils
- Rewrite quota-windows.ts as a 29-line thin aggregator with zero provider knowledge
- Wire getQuotaWindows into adapter registry for claude-local and codex-local
- Add 47 unit tests covering toPercent, secondsToWindowLabel, WHAM normalization,
readClaudeToken, readCodexToken, fetchClaudeQuota, fetchCodexQuota, fetchWithTimeout
- Add 8 unit tests covering parseDateRange validation and byProvider pro-rata math
Adding a third provider now requires only touching that provider's adapter.
Dotta