[waleedlatif1]

Summary

• Entra directory sync: a user deleted after the users phase (Graph 404 Request_ResourceNotFound on its directReports) reads as having no reports instead of discarding the whole staged run. Any other 404 still fails as before
• Entra directory sync: a person read under a second manager during a multi-slice read takes the latest read instead of failing the run. Graph allows one manager, so a second claim is a reorg mid-read; publication's cycle check still guards the final graph
• People profile and org chart keep loaded data when a background refresh fails with a 5xx or network error, so an open Edit modal no longer unmounts and loses its draft. 401/403/404 still replace the view as before
• People directory no longer requests the next page while showing the previous filter's placeholder rows
• Search Assistant refuses citation URLs with userinfo (https://trusted@other/) now that web results can cite
• Search clean-entry check uses toString() instead of URLSearchParams.size, which older Safari lacks
• Cleanup: checkpoint(..., { lockProfiles }) instead of a misnamed step arg, shared RELEASED, chunkArray, flattenFieldErrors, MCP TSDoc
• CLI, lint-test and run-hook cleanups from the same audit landed in improvement(cleanup): tighten the agent-harness changes from the latest audit #9650; this branch takes staging's versions of those files

Type of Change

• Bug fix

Testing

• lib/people/sync/stage.integration.ts and lib/people/sync/large-directory.integration.ts against real Postgres: 12/12 pass; each new case fails with its fix reverted
• resolve-citations.test.ts, retrieval-citations.test.ts, knowledge/application/chat.test.ts, slack-search/assistant-stream.test.ts: 71/71 pass; the credentialed-URL case fails with the fix reverted
• biome on changed files; type-check, audits and the full suite in CI

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing (new tests pass the test-audit authoring gate)
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
docs	Ready	Preview	Nov 4, 2026 1:42am UTC

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge; no new actionable issues remain.

Summary

Makes Entra sync tolerate users deleted mid-read and reporting lines that change between reads.

• Entra sync finishes snapshots when users change mid-read.
• People profiles and charts stay visible through temporary refresh errors.
• Citations skip web addresses that contain login details.
• The people list waits for its current filter before loading more.
• Search Assistant recognizes query strings on older Safari.
• Sync helpers and error notes use clearer shared names.

The previous unnumbered test finding is addressed: one call now includes competing manager claims and checks the intended winner. waleedlatif1 confirmed that same-batch claims deliberately choose the lowest manager ID, while later batches replace earlier assignments.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  Users[Read users into private stage] --> Reports[Read reporting lines]
  Reports --> Deleted{User deleted mid-read?}
  Deleted -->|Yes| Complete[Finish that user's reports read]
  Deleted -->|No| Claims[Save report claims]
  Claims --> Complete
  Complete --> More{More pages remain?}
  More -->|Yes| Reports
  More -->|No| Check[Check finished graph for cycles]
  Check -->|Valid| Publish[Publish together]
  Check -->|Cycle| Reject[Keep previous directory]

Loading

_{Reviews (3) · Last reviewed commit: "test(people): pin the same-batch manager..." · Reviewed by Greptile}

[cubic-dev-ai]

All reported issues were addressed across 20 files

<sub>Reply with feedback, questions, or to request a fix.

Turn on auto-fix | Re-trigger cubic</sub>

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 14 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}