Fix scheduled trigger invocations being skipped#2777
Merged
shagun-singh-inkeep merged 2 commits intomainfrom Mar 19, 2026
Merged
Fix scheduled trigger invocations being skipped#2777shagun-singh-inkeep merged 2 commits intomainfrom
shagun-singh-inkeep merged 2 commits intomainfrom
Conversation
…d without changing the next execution time
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
🦋 Changeset detectedLatest commit: 1e7b737 The changes in this PR will be included in the next version bump. This PR includes changesets to release 10 packages
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
Contributor
|
TL;DR — When a scheduled trigger is edited without changing its next execution time, the restarted workflow finds a cancelled invocation via idempotency and incorrectly skips it. This fix detects cancelled idempotent invocations, resets future ones to pending so they still fire, and chains past ones to the next iteration instead of silently dropping them. Key changes
Summary ｜ 4 files ｜ 2 commits ｜ base: Cancelled idempotent invocation recovery in
|
Contributor
There was a problem hiding this comment.
Solid fix that addresses a real gap in the idempotency logic for restarted workflows. The core idea — detecting cancelled invocations returned by the idempotent insert and either resetting them (future) or skipping to the next iteration (past) — is sound and well-structured.
One potential issue: the one-time trigger early return at line 219 may shadow this fix for one-time triggers.
| @@ -223,6 +224,55 @@ async function _scheduledTriggerRunnerWorkflow(payload: ScheduledTriggerRunnerPa | |||
| }); | |||
| return { status: 'already_executed', invocationId: invocation.id }; | |||
Contributor
There was a problem hiding this comment.
The condition on line 219 (isOneTime && result.alreadyExists && invocation.status !== 'pending') returns 'already_executed' for all non-pending statuses, including 'cancelled'. This means the new cancelled-invocation handling at line 228 is unreachable for one-time triggers.
If a one-time trigger's invocation is cancelled (e.g., trigger edited with runAt resent in the PATCH body, triggering scheduleChanged), the workflow will incorrectly report 'already_executed' instead of resetting to pending.
Consider narrowing the early return to exclude 'cancelled':
if (isOneTime && result.alreadyExists && invocation.status !== 'pending' && invocation.status !== 'cancelled') {Or — if one-time triggers should genuinely never recover from cancellation — add a comment explaining why.
| invocationId: params.invocationId, | ||
| data: { | ||
| status: 'pending', | ||
| attemptNumber: 1, |
Contributor
There was a problem hiding this comment.
Resetting attemptNumber: 1 is correct for a fresh start, but worth verifying that updateScheduledTriggerInvocationStatus accepts partial data here — specifically that it won't null out fields like startedAt or completedAt from the prior cancelled state. If those timestamp columns retain stale values from the cancelled run, downstream reporting/UI could be misleading.
| scheduledTriggerId, | ||
| invocationId: invocation.id, | ||
| }); | ||
| invocation = { ...invocation, status: 'pending' }; |
Contributor
There was a problem hiding this comment.
The local invocation is reassigned with status: 'pending', but the DB record was updated asynchronously in the step above. Between the reset and the sleep at line 293, checkInvocationCancelledStep re-fetches from the DB (line 313), so correctness isn't affected — but invocation.attemptNumber here retains the old value. This is fine because attemptNumber is re-read from the DB at line 350, but it's worth noting in case future code uses the local copy.
Contributor
There was a problem hiding this comment.
PR Review Summary
(4) Total Issues | Risk: Medium
🟠⚠️ Major (2) 🟠⚠️
Inline Comments:
- 🟠 Major:
scheduledTriggerRunner.ts:219Logic bug - one-time triggers with cancelled invocations won't be reset to pending - 🟠 Major:
scheduledTriggerSteps.ts:506Missing status guard against concurrent modifications in resetInvocationToPendingStep
🟡 Minor (2) 🟡
🟡 1) scheduledTriggerRunner.ts Missing test coverage for new code paths
Issue: The PR introduces three distinct code paths for handling cancelled invocations (lines 228-275) but includes no unit tests. The existing test suite for scheduled triggers only covers CRUD/API-level operations, not workflow execution logic.
Why: This is a bug fix for a production issue where trigger invocations were being skipped. Without tests, the same bug class could regress. The time-based branching logic combined with trigger type creates multiple code paths that could fail silently.
Fix: Add unit tests covering:
- Reset cancelled invocation to pending when scheduled time is in future
- Chain to next iteration for past cancelled cron trigger
- Return
skipped_cancelledfor past cancelled one-time trigger - Error propagation when chaining fails after cancelled invocation
Refs:
Inline Comments:
- 🟡 Minor:
scheduledTriggerRunner.ts:238-244Missing try-catch for error context on resetInvocationToPendingStep call
💭 Consider (2) 💭
💭 1) scheduledTriggerRunner.ts:233-237 Add tenant/project/agent context to logs
Issue: The log statements for cancelled invocation handling omit tenantId, projectId, and agentId.
Why: Other logs in the file (e.g., line 151-158) include full scope context. Consistency aids debugging.
Fix: Add scope fields to log payloads at lines 233-237 and 247-251.
💭 2) scheduledTriggerRunner.ts:229-230 Consider atomic time check
Issue: The time comparison uses Date.now() in application code, then the database update happens separately. In a distributed environment, clock drift or processing delays could cause decisions to become stale.
Why: This is a defense-in-depth concern. The status guard (Major #2) is more critical since the workflow framework handles retries.
Fix: If needed, move time check into the database WHERE clause for atomic decision-making.
🚫 REQUEST CHANGES
Summary: The fix for handling cancelled invocations is a good approach, but there's a logic bug that causes one-time triggers to still skip cancelled invocations (they return "already_executed" before reaching the new fix). The suggested one-line fix at line 219 should resolve this. Additionally, consider adding a status guard to resetInvocationToPendingStep to prevent race conditions with concurrent cancellations.
Discarded (3)
| Location | Issue | Reason Discarded |
|---|---|---|
scheduledTriggerSteps.ts:506 |
Missing error logging in step function | The workflow framework handles retries and the caller logs context. Adding logging here would be redundant with the suggested try-catch in the caller. |
scheduledTriggerRunner.ts:229 |
TOCTOU race condition on time check | Low severity in practice since workflow framework retries. The status guard is the more important fix. |
scheduledTriggerRunner.ts:245 |
Local invocation mutation not verified | The mutation is straightforward and the workflow proceeds correctly. Test coverage is covered by Major finding. |
Reviewers (4)
| Reviewer | Returned | Main Findings | Consider | Inline Comments | Discarded |
|---|---|---|---|---|---|
pr-review-standards |
1 | 0 | 0 | 1 | 0 |
pr-review-errors |
2 | 0 | 0 | 1 | 1 |
pr-review-tests |
3 | 1 | 0 | 0 | 1 |
pr-review-sre |
5 | 0 | 2 | 1 | 1 |
| Total | 11 | 1 | 2 | 3 | 3 |
| }) { | ||
| 'use step'; | ||
|
|
||
| await updateScheduledTriggerInvocationStatus(runDbClient)({ |
Contributor
There was a problem hiding this comment.
🟠 MAJOR Missing status guard against concurrent modifications
Issue: This update unconditionally sets status to 'pending' without checking the current status. The existing markScheduledTriggerInvocationCompleted and markScheduledTriggerInvocationFailed functions both include ne(status, 'cancelled') guards, but this function lacks similar protection.
Why: If another workflow instance or user cancellation changes the status between the idempotency check (line 228 in runner) and this update, the reset could override a legitimate state transition. This could cause a race condition where a user's cancellation is silently ignored.
Fix: The underlying updateScheduledTriggerInvocationStatus DAL function is generic, so the guard should be added at the application layer. Consider either:
- Add a
WHERE status = 'cancelled'condition via a dedicated DAL function - Check the returned row count and handle the case where no update occurred
Refs:
Comment on lines
+238
to
+244
| await resetInvocationToPendingStep({ | ||
| tenantId, | ||
| projectId, | ||
| agentId, | ||
| scheduledTriggerId, | ||
| invocationId: invocation.id, | ||
| }); |
Contributor
There was a problem hiding this comment.
🟡 Minor Missing try-catch for error context
Issue: This call to resetInvocationToPendingStep lacks a try-catch wrapper, unlike startNextIterationStep at lines 254-270 which properly catches errors, logs context, and re-throws.
Why: If resetInvocationToPendingStep fails, the workflow will error out without a helpful log message indicating which operation failed. The workflow framework will retry, but debugging will be harder.
Fix: Add consistent error handling:
try {
await resetInvocationToPendingStep({
tenantId,
projectId,
agentId,
scheduledTriggerId,
invocationId: invocation.id,
});
} catch (err) {
await logStep('Failed to reset cancelled invocation to pending', {
scheduledTriggerId,
invocationId: invocation.id,
error: err instanceof Error ? err.message : String(err),
});
throw err;
}
Contributor
There was a problem hiding this comment.
PR Review Summary
(0) Total Issues | Risk: Low
This is a delta review of commit 1e7b73745 ("claude comments") which addresses all major feedback from the prior review.
✅ Prior Feedback Addressed
The author has addressed all major and minor issues from the previous review:
| Prior Issue | Status |
|---|---|
| 🟠 One-time trigger early return shadowing fix (line 219) | ✅ Fixed — condition now excludes cancelled status |
🟠 Missing status guard in resetInvocationToPendingStep |
✅ Fixed — new resetCancelledInvocationToPending DAL function with WHERE status = 'cancelled' |
| 🟡 Missing try-catch for error context | ✅ Fixed — proper try-catch added with contextual logging |
| 🟡 Concurrent modification handling | ✅ Fixed — returns early with concurrent_status_change if reset returns undefined |
💭 Consider (still open from prior review)
These were suggestions from the prior review that remain unaddressed but are non-blocking:
-
Missing test coverage — The new code paths (cancelled invocation handling) have no unit tests. Consider adding tests for:
- Reset cancelled invocation to pending when scheduled time is in future
- Chain to next iteration for past cancelled cron trigger
- Return
skipped_cancelledfor past cancelled one-time trigger
-
Log context consistency — The log statements at lines 238-242, 253-257, 262-264, and 270-274 omit
tenantId,projectId, andagentIdwhich are included in other logs in the file (e.g., line 151-158).
Delta Review Summary
The delta changes are well-implemented:
- DAL layer (
resetCancelledInvocationToPending): Properly scoped WHERE clause withstatus = 'cancelled'guard, clears stale timestamps, returns undefined on no-op - Step function: Returns result to caller, appropriate log levels (info/warn)
- Workflow: Proper try-catch pattern consistent with
startNextIterationStep, handles undefined result with early return
No new bugs, correctness issues, or regressions introduced.
✅ APPROVE
Summary: All blocking issues from the prior review have been addressed. The fix correctly handles cancelled idempotent invocations by either resetting them to pending (future) or skipping to the next iteration (past), with proper concurrency guards. The remaining suggestions (tests, log context) are non-blocking quality improvements that can be addressed in a follow-up.
Reviewers (2)
| Reviewer | Returned | Main Findings | Consider | Inline Comments | Discarded |
|---|---|---|---|---|---|
pr-review-standards |
0 | 0 | 0 | 0 | 0 |
pr-review-errors |
0 | 0 | 0 | 0 | 0 |
| Total | 0 | 0 | 0 | 0 | 0 |
Note: Delta review — reviewers confirmed no new issues in the changes since last review.
dimaMachina
pushed a commit
that referenced
this pull request
Mar 20, 2026
* Fix scheduled trigger invocations being skipped when trigger is edited without changing the next execution time * claude comments
github-merge-queue Bot
pushed a commit
that referenced
this pull request
Mar 30, 2026
* skill generator * polish skill generator * skills tests * upd * upd * generation.test wip * add generation.test * tree node * skill page * skill loader * skill loader refactor * skill loader * move skills sidebar to layout * use pure monaco-editor component since we can have different file extension * add shadcn context menu component * format context menu * skills files and edit pages * dry * update layout * add docs * add a changeset * redirect to first skill * skill files utils * skill selector * upd treenode * skill files * skill file editor * delete skill confirmation * add skill files actions * skills data * rm * up skills route * upd * upd * better project error message on dev * types * skill files * skill loader * format * project test * entities * project full tests * upd introspect * upd cliiii * nested skills tests * remove edit page * remove edit page * update files page * upd * upd file editor * add SkillFileInsertSchema * superRefine * add transform * rm some cases in superRefine * use pipe * use pipe * upd skill loader * validation skills * upd * rm * upd * data access tests * skills db changes * add * skill files * upd * upd * upd skill update * SkillUpdateSchema has required files * upd skills manage * upd * upd layout and page * style: auto-format with biome * move empty state comp to page * upd schemas * update schemas * move to with-sidebar * polish * upd * upd skill generator * Make webhooks docs user friendly (#2752) * shaping a2a webhooks page * moved triggers to visual builder * vb webhooks wip * numbered TOC steps * added step circles * indented toc steps more * added newsletter signup to docs * added share feedback button * moved newsletter subscribe route to agent docs * subscribe confirm polish * improved spacing * improved spacing * added high quality images * added verification step * Sync lockfile after rebase * Use tag reference for pullfrog action instead of pinned SHA (#2757) The action is actively under development, so referencing the v0 tag allows picking up updates automatically. https://claude.ai/code/session_01QZyvEs97scVf1ahTG8C1rV Co-authored-by: Claude <[email protected]> * ci: provision PR preview environments in Railway (#2681) * ci: add preview env diagnostics * ci: probe preview env schema before deploy * ci: probe preview env schema before deploy * ci: harden preview api env defaults * ci: attach git metadata to preview deploys * ci: harden preview workflow operations * ci: broaden preview log redaction * ci: extract preview workflow scripts * ci: harden preview script extraction * fix(ci): correct Playwright cache restore-key prefix mismatch (#2760) The restore-keys used `${{ runner.os }}-playwright-` but primary keys used `playwright-${{ runner.os }}-`, so the prefix never matched on cache miss, forcing a full browser download (~8.5 min) instead of a cache restore (~13 sec). Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> * fix(ci): replace full git clone with shallow checkout in CI job (#2761) Remove fetch-depth: 0 from the ci job's checkout step, which cloned the entire git history (1.5-5 min overhead). Only the OpenAPI change detection step needs the base branch ref, so fetch it on-demand with --depth=1. Also switches the diff from three-dot merge-base syntax to a two-dot pathspec-filtered diff against the fetched base ref, which works correctly with shallow clones. Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> * format * rm migration * add new migrations * validation for skill is ok * move empty state to page * delete skill * delete skill revalidate path * move skills schemas to own file * upd * upd * upd * upd * upd * upd * upd * upd * upd * upd * more typecheck fixes * more typecheck fixes * fix * fix isRequired * f1x * move skill sidebar * refactor skill sidebar * add collapse file tree button * upd * upd * upd * deleteSkillFile * upd * deleteSkillFile * fileId * fileId * upd schemas * DeleteSkillFileConfirmation * updateSkillFile * rm simplematter from sdk * Get Skill File * getSkillFileById * add new skill file page * update skill file editor * format * Create Skill File * upd * createSkillFileAction * createSkillFileById * fix: Make OpenTelemetry startup idempotent (#2684) * fix: Make OpenTelemetry startup idempotent * fix: Re-export defaultSDK and cache NodeSDK instance on globalThis Restores the export on defaultSDK to avoid breaking the create-agents-template subpath import. Moves the new NodeSDK() construction behind a globalThis guard (getOrCreateSDK) so repeated Vite HMR module evaluations reuse the same instance instead of leaking fresh SDK objects. Co-authored-by: mike-inkeep <[email protected]> Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix(template): use idempotent startOpenTelemetrySDK() in instrumentation * fix: guard all OTel singletons behind globalThis for HMR idempotency - Cache otlpExporter, batchProcessor, resource, instrumentations, spanProcessors, contextManager, and propagator on globalThis via Symbol keys and getOrCreate* helpers so HMR re-evaluation reuses existing instances instead of leaking new ones - Make OtelGlobal type strict with per-key types, eliminating the loose `boolean | NodeSDK` union and the `as NodeSDK` cast - Add logger.debug in the MetricReader catch block to distinguish clean idempotency from error-recovery idempotency - Remove defaultSDK export (now module-private) since all consumers use startOpenTelemetrySDK() instead * Fix type errors * Simplify to just suppress the error since it's not an issue in prod, only local * Limit to dev mode * Add changeset for OTel HMR idempotency fix Co-authored-by: Dimitri POSTOLOV <[email protected]> Co-Authored-By: Claude Opus 4.6 <[email protected]> --------- Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: mike-inkeep <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: inkeep[bot] <257615677+inkeep[bot]@users.noreply.github.com> Co-authored-by: Dimitri POSTOLOV <[email protected]> * Fix scheduled trigger invocations being skipped (#2777) * Fix scheduled trigger invocations being skipped when trigger is edited without changing the next execution time * claude comments * adding app id (#2779) * Update pullfrog to latest SHA and add daily dependabot group for high-frequency deps (#2780) * Update pullfrog to latest SHA and add daily dependabot group for high-frequency deps - Update pullfrog from v0.0.178 SHA to v0.0.181 SHA (30d68e5) to stay on commit-pinned references for security (action has write permissions + 9 API keys) - Split dependabot github-actions config into a "high-frequency" group for pullfrog with daily schedule, so SHA pins get updated automatically - This supersedes PR #2757's approach of moving to mutable tag references https://claude.ai/code/session_01D3ZGYHG8VhsZwqjjXqy2Ap * Split dependabot github-actions into daily (pullfrog) and monthly (rest) Separate into two ecosystem entries so pullfrog gets daily SHA updates while other GitHub Actions stay on a monthly cadence. https://claude.ai/code/session_01D3ZGYHG8VhsZwqjjXqy2Ap * Fix invalid dependabot config: merge duplicate github-actions entries Dependabot disallows duplicate ecosystem+directory pairs. Use a single entry with two groups instead: high-frequency (pullfrog) and github-actions (everything else via exclude-patterns). https://claude.ai/code/session_01D3ZGYHG8VhsZwqjjXqy2Ap --------- Co-authored-by: Claude <[email protected]> * ci: seed preview auth in PR previews (#2775) * ci: bootstrap preview auth * ci: require secure preview auth config * ci: recover preview auth runtime vars * ci: install railway in preview bootstrap * ci: provision preview db tcp proxies * ci: proxy preview spicedb bootstrap * ci: harden preview retry and error logging --------- Co-authored-by: Andrew Mikofalvy <[email protected]> * Fix scopes placeholder to show correct Nango format (#2784) * Fix misleading scopes placeholder in credential form The Nango API validates scopes against a strict comma-separated pattern with no spaces. Updated placeholder and help text to show the correct format and prevent 400 errors when users enter multiple scopes. Made-with: Cursor * Add changeset for scopes placeholder fix Made-with: Cursor * fix(manage-ui): fix URL validation bypass and permission guard in credential provider setup (#2776) * fix(manage-ui): fix URL validation bypass and permission guard in credential provider setup Reorder Zod schema construction so custom validators (e.g. URL protocol allowlist) are chained after required/optional base schema instead of being overwritten. Move all React hooks above the canEdit early-return guard to satisfy Rules of Hooks, with canEdit checks inside hook bodies. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * fix(manage-ui): add server-side URL protocol validation in buildCredentialsPayload Validate app_link against HTTP/HTTPS allowlist in the server action to prevent bypassing client-side form validation. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Update agents-manage-ui/src/components/credentials/views/generic-auth-form.tsx Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * fix err --------- Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * feat(pdf): Support PDF attachments (#2709) * feat(pdf): Support PDF attachments * Add tests and other review feedback * Fix doc * More renaming and cleanup * refactor: extract Vercel content part schemas to types/chat.ts for reuse Move inline Zod schemas from chatDataStream.ts and message-parts.ts into types/chat.ts as shared, exported schemas. This eliminates duplicate definitions and makes schema management easier. Co-authored-by: Andrew Mikofalvy <[email protected]> Co-Authored-By: Claude Opus 4.6 <[email protected]> --------- Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Andrew Mikofalvy <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> * feat: Composio connected account ID pinning (#2783) * feat: Composio connected account ID pinning Pin connected_account_id to Composio MCP URLs to prevent cross-project credential leakage. Implements "both or none" policy — user_id and connected_account_id are injected together or not at all. - Add ComposioCredentialStore for credential lifecycle management - Update AgentMcpManager and discoverToolsFromServer with pinning logic - Mark Composio tools without connectedAccountId as needs_auth - Add generic disconnect credential UI (works for all credential types) - Store authScheme in credential retrievalParams for display - Update OAuth login flow to create credential references post-connect - Add unit tests for new credential store, composio client, and pinning Made-with: Cursor * feedback * fix test * Version Packages (#2778) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * Add back link to projects sidebar, add org settings link to user drop… (#2787) * Add back link to projects sidebar, add org settings link to user dropdown, adjust sidebar highlight color in dark mode * Apply suggestion from @claude[bot] Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * Fix bad claude formatting --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * fix: return FileUIPart-compliant file parts from /run conversations endpoint (#2782) * fix: return Vercel AI SDK FileUIPart-compliant file parts from /run conversations endpoint - Resolve blob:// URIs to proxy HTTP URLs via resolveMessagesListBlobUris() - Reshape file parts from { data, metadata.mimeType } to { url, mediaType, filename? } - Matches Vercel AI SDK FileUIPart spec for useChat() compatibility Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Skip malformed file parts --------- Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> Co-authored-by: Mike Rashkovsky <[email protected]> * fix: treat load_skill as internal tool to suppress false-positive Sentry errors (#2756) * fix: provide relationshipId for load_skill tool calls in graph events * fix: treat load_skill as internal tool, suppress chat/graph streaming events * fix for fetch trace (#2791) * fix for fetch trace * fix for fetch trace * Fix empty breadcrumb on `/[tenantId]/profile` page and replace prop-drilled permission flags (`readOnly`, `canEdit`, `canUse`) with direct hook call `useProjectPermissionsQuery()` (#2792) * upd * upd * format * format * format * format * format * format * format * format * format * fix review * fix breadcrumb on profile page * Apply suggestions from code review Co-authored-by: Dimitri POSTOLOV <[email protected]> * Update agents-manage-ui/src/lib/api/projects.ts Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * Update agents-manage-ui/src/app/[tenantId]/profile/layout.tsx Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * style: auto-format with biome * fix review --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * fix(manage-ui): fix user-scoped MCP credential card not refreshing after connect/disconnect (#2794) Fetch user-scoped credential server-side in page.tsx (matching the project-scoped pattern) instead of via a client-side React Query hook. This ensures router.refresh() after OAuth connect or credential delete re-fetches the credential data, so the "Your Connection" card updates without a manual page refresh. Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> * reuse `useProjectsQuery` instead of `fetchProjectsAction` in `useEffect` (#2793) * reuse `useProjectsQuery` instead of `fetchProjectsAction` in `useEffect` * format * upd * fix lint * Create little-hounds-battle.md * upd * upd skill file editor * polish skill editor like in github * remove canEdit * upd * move skill metadata under collapsible advanced section * reuse DeleteConfirmation * upd skill file editor * upd skill file editor2 * add useInitialCollapsedSidebar * add useInitialCollapsedSidebar * upd skill file editor * rm * // Avoid including metadata in the frontmatter when it's null * fetchSkillFile and createSkillFile * refactor skill breadcrumb * format * polish * upd * skills integration tests * fix validation tests * update skill form * upd api skills in manage ui * upd entities * partial * fix skill loader test * chore: update OpenAPI snapshot * polish skill file editor * upd core skills tests * upd core skills tests * add SkillCreateDataSchema * update skills data manage * remove redundant * lint * lint * typecheck * typecheck * typecheck * knip * lint * rollback skill modals * make modal opens in skill selector * fix typecheck * this should fix cypress * fix sdk tests * split permissions call * add folder feature * findNodeByPath * SkillDirectoryBrowser * upd * polish * fix * fix edge case metadata validation * fix * fix cli test * format * upd * upd * upd * chore: update OpenAPI snapshot * fix skill generator * add button group * connect submit logic with extension select * polish * update skill generator tests * update generation test * polish skill generator * format * format * fixes for tests * typecheck * fix review * format * new migration * upd * rm migration * add migrations * fix migration and add * rm outdated * Apply suggestions from code review Co-authored-by: pullfrog[bot] <226033991+pullfrog[bot]@users.noreply.github.com> Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * polish * Add detailed changeset for nested skill files feature Co-authored-by: Dimitri POSTOLOV <[email protected]> * fix typecheck --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Gaurav Varma <[email protected]> Co-authored-by: Andrew Mikofalvy <[email protected]> Co-authored-by: Claude <[email protected]> Co-authored-by: Varun Varahabhotla <[email protected]> Co-authored-by: mike-inkeep <[email protected]> Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: mike-inkeep <[email protected]> Co-authored-by: inkeep[bot] <257615677+inkeep[bot]@users.noreply.github.com> Co-authored-by: Dimitri POSTOLOV <[email protected]> Co-authored-by: shagun-singh-inkeep <[email protected]> Co-authored-by: omar-inkeep <[email protected]> Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> Co-authored-by: Andrew Mikofalvy <[email protected]> Co-authored-by: inkeep-internal-ci[bot] <259778081+inkeep-internal-ci[bot]@users.noreply.github.com> Co-authored-by: sarah <[email protected]> Co-authored-by: Abraham <[email protected]> Co-authored-by: pullfrog[bot] <226033991+pullfrog[bot]@users.noreply.github.com>
tim-inkeep
pushed a commit
that referenced
this pull request
Mar 31, 2026
* skill generator * polish skill generator * skills tests * upd * upd * generation.test wip * add generation.test * tree node * skill page * skill loader * skill loader refactor * skill loader * move skills sidebar to layout * use pure monaco-editor component since we can have different file extension * add shadcn context menu component * format context menu * skills files and edit pages * dry * update layout * add docs * add a changeset * redirect to first skill * skill files utils * skill selector * upd treenode * skill files * skill file editor * delete skill confirmation * add skill files actions * skills data * rm * up skills route * upd * upd * better project error message on dev * types * skill files * skill loader * format * project test * entities * project full tests * upd introspect * upd cliiii * nested skills tests * remove edit page * remove edit page * update files page * upd * upd file editor * add SkillFileInsertSchema * superRefine * add transform * rm some cases in superRefine * use pipe * use pipe * upd skill loader * validation skills * upd * rm * upd * data access tests * skills db changes * add * skill files * upd * upd * upd skill update * SkillUpdateSchema has required files * upd skills manage * upd * upd layout and page * style: auto-format with biome * move empty state comp to page * upd schemas * update schemas * move to with-sidebar * polish * upd * upd skill generator * Make webhooks docs user friendly (#2752) * shaping a2a webhooks page * moved triggers to visual builder * vb webhooks wip * numbered TOC steps * added step circles * indented toc steps more * added newsletter signup to docs * added share feedback button * moved newsletter subscribe route to agent docs * subscribe confirm polish * improved spacing * improved spacing * added high quality images * added verification step * Sync lockfile after rebase * Use tag reference for pullfrog action instead of pinned SHA (#2757) The action is actively under development, so referencing the v0 tag allows picking up updates automatically. https://claude.ai/code/session_01QZyvEs97scVf1ahTG8C1rV Co-authored-by: Claude <[email protected]> * ci: provision PR preview environments in Railway (#2681) * ci: add preview env diagnostics * ci: probe preview env schema before deploy * ci: probe preview env schema before deploy * ci: harden preview api env defaults * ci: attach git metadata to preview deploys * ci: harden preview workflow operations * ci: broaden preview log redaction * ci: extract preview workflow scripts * ci: harden preview script extraction * fix(ci): correct Playwright cache restore-key prefix mismatch (#2760) The restore-keys used `${{ runner.os }}-playwright-` but primary keys used `playwright-${{ runner.os }}-`, so the prefix never matched on cache miss, forcing a full browser download (~8.5 min) instead of a cache restore (~13 sec). Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> * fix(ci): replace full git clone with shallow checkout in CI job (#2761) Remove fetch-depth: 0 from the ci job's checkout step, which cloned the entire git history (1.5-5 min overhead). Only the OpenAPI change detection step needs the base branch ref, so fetch it on-demand with --depth=1. Also switches the diff from three-dot merge-base syntax to a two-dot pathspec-filtered diff against the fetched base ref, which works correctly with shallow clones. Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> * format * rm migration * add new migrations * validation for skill is ok * move empty state to page * delete skill * delete skill revalidate path * move skills schemas to own file * upd * upd * upd * upd * upd * upd * upd * upd * upd * upd * more typecheck fixes * more typecheck fixes * fix * fix isRequired * f1x * move skill sidebar * refactor skill sidebar * add collapse file tree button * upd * upd * upd * deleteSkillFile * upd * deleteSkillFile * fileId * fileId * upd schemas * DeleteSkillFileConfirmation * updateSkillFile * rm simplematter from sdk * Get Skill File * getSkillFileById * add new skill file page * update skill file editor * format * Create Skill File * upd * createSkillFileAction * createSkillFileById * fix: Make OpenTelemetry startup idempotent (#2684) * fix: Make OpenTelemetry startup idempotent * fix: Re-export defaultSDK and cache NodeSDK instance on globalThis Restores the export on defaultSDK to avoid breaking the create-agents-template subpath import. Moves the new NodeSDK() construction behind a globalThis guard (getOrCreateSDK) so repeated Vite HMR module evaluations reuse the same instance instead of leaking fresh SDK objects. Co-authored-by: mike-inkeep <[email protected]> Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix(template): use idempotent startOpenTelemetrySDK() in instrumentation * fix: guard all OTel singletons behind globalThis for HMR idempotency - Cache otlpExporter, batchProcessor, resource, instrumentations, spanProcessors, contextManager, and propagator on globalThis via Symbol keys and getOrCreate* helpers so HMR re-evaluation reuses existing instances instead of leaking new ones - Make OtelGlobal type strict with per-key types, eliminating the loose `boolean | NodeSDK` union and the `as NodeSDK` cast - Add logger.debug in the MetricReader catch block to distinguish clean idempotency from error-recovery idempotency - Remove defaultSDK export (now module-private) since all consumers use startOpenTelemetrySDK() instead * Fix type errors * Simplify to just suppress the error since it's not an issue in prod, only local * Limit to dev mode * Add changeset for OTel HMR idempotency fix Co-authored-by: Dimitri POSTOLOV <[email protected]> Co-Authored-By: Claude Opus 4.6 <[email protected]> --------- Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: mike-inkeep <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: inkeep[bot] <257615677+inkeep[bot]@users.noreply.github.com> Co-authored-by: Dimitri POSTOLOV <[email protected]> * Fix scheduled trigger invocations being skipped (#2777) * Fix scheduled trigger invocations being skipped when trigger is edited without changing the next execution time * claude comments * adding app id (#2779) * Update pullfrog to latest SHA and add daily dependabot group for high-frequency deps (#2780) * Update pullfrog to latest SHA and add daily dependabot group for high-frequency deps - Update pullfrog from v0.0.178 SHA to v0.0.181 SHA (30d68e5) to stay on commit-pinned references for security (action has write permissions + 9 API keys) - Split dependabot github-actions config into a "high-frequency" group for pullfrog with daily schedule, so SHA pins get updated automatically - This supersedes PR #2757's approach of moving to mutable tag references https://claude.ai/code/session_01D3ZGYHG8VhsZwqjjXqy2Ap * Split dependabot github-actions into daily (pullfrog) and monthly (rest) Separate into two ecosystem entries so pullfrog gets daily SHA updates while other GitHub Actions stay on a monthly cadence. https://claude.ai/code/session_01D3ZGYHG8VhsZwqjjXqy2Ap * Fix invalid dependabot config: merge duplicate github-actions entries Dependabot disallows duplicate ecosystem+directory pairs. Use a single entry with two groups instead: high-frequency (pullfrog) and github-actions (everything else via exclude-patterns). https://claude.ai/code/session_01D3ZGYHG8VhsZwqjjXqy2Ap --------- Co-authored-by: Claude <[email protected]> * ci: seed preview auth in PR previews (#2775) * ci: bootstrap preview auth * ci: require secure preview auth config * ci: recover preview auth runtime vars * ci: install railway in preview bootstrap * ci: provision preview db tcp proxies * ci: proxy preview spicedb bootstrap * ci: harden preview retry and error logging --------- Co-authored-by: Andrew Mikofalvy <[email protected]> * Fix scopes placeholder to show correct Nango format (#2784) * Fix misleading scopes placeholder in credential form The Nango API validates scopes against a strict comma-separated pattern with no spaces. Updated placeholder and help text to show the correct format and prevent 400 errors when users enter multiple scopes. Made-with: Cursor * Add changeset for scopes placeholder fix Made-with: Cursor * fix(manage-ui): fix URL validation bypass and permission guard in credential provider setup (#2776) * fix(manage-ui): fix URL validation bypass and permission guard in credential provider setup Reorder Zod schema construction so custom validators (e.g. URL protocol allowlist) are chained after required/optional base schema instead of being overwritten. Move all React hooks above the canEdit early-return guard to satisfy Rules of Hooks, with canEdit checks inside hook bodies. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * fix(manage-ui): add server-side URL protocol validation in buildCredentialsPayload Validate app_link against HTTP/HTTPS allowlist in the server action to prevent bypassing client-side form validation. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Update agents-manage-ui/src/components/credentials/views/generic-auth-form.tsx Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * fix err --------- Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * feat(pdf): Support PDF attachments (#2709) * feat(pdf): Support PDF attachments * Add tests and other review feedback * Fix doc * More renaming and cleanup * refactor: extract Vercel content part schemas to types/chat.ts for reuse Move inline Zod schemas from chatDataStream.ts and message-parts.ts into types/chat.ts as shared, exported schemas. This eliminates duplicate definitions and makes schema management easier. Co-authored-by: Andrew Mikofalvy <[email protected]> Co-Authored-By: Claude Opus 4.6 <[email protected]> --------- Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: Andrew Mikofalvy <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> * feat: Composio connected account ID pinning (#2783) * feat: Composio connected account ID pinning Pin connected_account_id to Composio MCP URLs to prevent cross-project credential leakage. Implements "both or none" policy — user_id and connected_account_id are injected together or not at all. - Add ComposioCredentialStore for credential lifecycle management - Update AgentMcpManager and discoverToolsFromServer with pinning logic - Mark Composio tools without connectedAccountId as needs_auth - Add generic disconnect credential UI (works for all credential types) - Store authScheme in credential retrievalParams for display - Update OAuth login flow to create credential references post-connect - Add unit tests for new credential store, composio client, and pinning Made-with: Cursor * feedback * fix test * Version Packages (#2778) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * Add back link to projects sidebar, add org settings link to user drop… (#2787) * Add back link to projects sidebar, add org settings link to user dropdown, adjust sidebar highlight color in dark mode * Apply suggestion from @claude[bot] Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * Fix bad claude formatting --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * fix: return FileUIPart-compliant file parts from /run conversations endpoint (#2782) * fix: return Vercel AI SDK FileUIPart-compliant file parts from /run conversations endpoint - Resolve blob:// URIs to proxy HTTP URLs via resolveMessagesListBlobUris() - Reshape file parts from { data, metadata.mimeType } to { url, mediaType, filename? } - Matches Vercel AI SDK FileUIPart spec for useChat() compatibility Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * Skip malformed file parts --------- Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> Co-authored-by: Mike Rashkovsky <[email protected]> * fix: treat load_skill as internal tool to suppress false-positive Sentry errors (#2756) * fix: provide relationshipId for load_skill tool calls in graph events * fix: treat load_skill as internal tool, suppress chat/graph streaming events * fix for fetch trace (#2791) * fix for fetch trace * fix for fetch trace * Fix empty breadcrumb on `/[tenantId]/profile` page and replace prop-drilled permission flags (`readOnly`, `canEdit`, `canUse`) with direct hook call `useProjectPermissionsQuery()` (#2792) * upd * upd * format * format * format * format * format * format * format * format * format * fix review * fix breadcrumb on profile page * Apply suggestions from code review Co-authored-by: Dimitri POSTOLOV <[email protected]> * Update agents-manage-ui/src/lib/api/projects.ts Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * Update agents-manage-ui/src/app/[tenantId]/profile/layout.tsx Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * style: auto-format with biome * fix review --------- Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * fix(manage-ui): fix user-scoped MCP credential card not refreshing after connect/disconnect (#2794) Fetch user-scoped credential server-side in page.tsx (matching the project-scoped pattern) instead of via a client-side React Query hook. This ensures router.refresh() after OAuth connect or credential delete re-fetches the credential data, so the "Your Connection" card updates without a manual page refresh. Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]> * reuse `useProjectsQuery` instead of `fetchProjectsAction` in `useEffect` (#2793) * reuse `useProjectsQuery` instead of `fetchProjectsAction` in `useEffect` * format * upd * fix lint * Create little-hounds-battle.md * upd * upd skill file editor * polish skill editor like in github * remove canEdit * upd * move skill metadata under collapsible advanced section * reuse DeleteConfirmation * upd skill file editor * upd skill file editor2 * add useInitialCollapsedSidebar * add useInitialCollapsedSidebar * upd skill file editor * rm * // Avoid including metadata in the frontmatter when it's null * fetchSkillFile and createSkillFile * refactor skill breadcrumb * format * polish * upd * skills integration tests * fix validation tests * update skill form * upd api skills in manage ui * upd entities * partial * fix skill loader test * chore: update OpenAPI snapshot * polish skill file editor * upd core skills tests * upd core skills tests * add SkillCreateDataSchema * update skills data manage * remove redundant * lint * lint * typecheck * typecheck * typecheck * knip * lint * rollback skill modals * make modal opens in skill selector * fix typecheck * this should fix cypress * fix sdk tests * split permissions call * add folder feature * findNodeByPath * SkillDirectoryBrowser * upd * polish * fix * fix edge case metadata validation * fix * fix cli test * format * upd * upd * upd * chore: update OpenAPI snapshot * fix skill generator * add button group * connect submit logic with extension select * polish * update skill generator tests * update generation test * polish skill generator * format * format * fixes for tests * typecheck * fix review * format * new migration * upd * rm migration * add migrations * fix migration and add * rm outdated * Apply suggestions from code review Co-authored-by: pullfrog[bot] <226033991+pullfrog[bot]@users.noreply.github.com> Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> * polish * Add detailed changeset for nested skill files feature Co-authored-by: Dimitri POSTOLOV <[email protected]> * fix typecheck --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Gaurav Varma <[email protected]> Co-authored-by: Andrew Mikofalvy <[email protected]> Co-authored-by: Claude <[email protected]> Co-authored-by: Varun Varahabhotla <[email protected]> Co-authored-by: mike-inkeep <[email protected]> Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com> Co-authored-by: mike-inkeep <[email protected]> Co-authored-by: inkeep[bot] <257615677+inkeep[bot]@users.noreply.github.com> Co-authored-by: Dimitri POSTOLOV <[email protected]> Co-authored-by: shagun-singh-inkeep <[email protected]> Co-authored-by: omar-inkeep <[email protected]> Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com> Co-authored-by: Andrew Mikofalvy <[email protected]> Co-authored-by: inkeep-internal-ci[bot] <259778081+inkeep-internal-ci[bot]@users.noreply.github.com> Co-authored-by: sarah <[email protected]> Co-authored-by: Abraham <[email protected]> Co-authored-by: pullfrog[bot] <226033991+pullfrog[bot]@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fix scheduled trigger invocations being skipped when trigger is edited without changing the next execution time