CedarCopilot wants to merge 15 commits into staging from refactor/document-store-retention
Live on prod, 1 of 6 surfaces working, 2 days left
Strict CloudWatch fallback saw 221 success-shaped log lines matching pipeline, analytics, quota, aop, but no tied operation was present, so Libra is not calling this working.
Strict CloudWatch fallback saw 211 success-shaped log lines matching auth, login, session, token, but no tied operation was present, so Libra is not calling this working.
Verifier did not provide positive runtime evidence tied to this intent, so Libra is keeping it inconclusive. Original assessment: OTEL fallback saw 6,648 prod spans for the repo/env since deploy with 12 error spans across 5 identified users.
Strict CloudWatch fallback saw 231 success-shaped log lines matching calendar, meeting, meetings, but no tied operation was present, so Libra is not calling this working.
Strict CloudWatch fallback saw 302 success-shaped log lines matching email, emails, gmail, mail, inbox, but no tied operation was present, so Libra is not calling this working.
Strict CloudWatch fallback saw 167 success-shaped log lines matching crm, hubspot, salesforce, deal, deals, but no tied operation was present, so Libra is not calling this working.
single_user
internal_only
Libra measured 1 of 6 production surfaces on prod. 3 surfaces had under 20 requests, so Libra has not judged them; 2 surfaces had no requests at all. Libra checks hourly for 3 days after each deploy.
A Recent deals widget joins the home rail, and it reads the CRM's filter and sort rather than inventing its own , so the deals on the home page are the deals the user just arranged, in that order.
That reuse is the change. The canvas's filter/sort config moves out of conversation-canvas/ into crm/utils/, which lets the /conversations sidebar hook drop 136 lines of duplicated variable-building; both surfaces now share one mapping that cannot disagree about an operator.
Alongside it: the home rail loses its bespoke machinery (use-home-teams-collapsed and the onboarding HomeWidgetsStep are gone, layout is one HomeLayoutDialog), and four leadership templates join the agent library , forecast-roll-up, team-coaching, release-feedback, competitor-intel.
| Gate | Result |
|---|---|
/review | Fixed: a date-operator cast hiding gt/lt/eq, six _-prefixed voided consts, a ! on a palette lookup, a widget querying before its stored rule arrived |
/thermo-review | Fixed: four new library templates missing from the marker-model test, dead setup-idle-* CSS, two comments asserting the opposite of the code |
| Greptile CLI | 1/5 → 4/5 over two iterations |
Greptile caught a real bug in a fix from the round before it: the cron assertion compared a bare expression against 0 7 * * 2 (America/Los_Angeles), so it could never pass , and it failed after the write landed, which would have aborted a production migration partway through.
Every job in ci.yml, run locally first. This surfaced failures that would otherwise have landed here red:
namespace: was never pinned on a subagent write. reconcileSubagentWrite carried an empty if (pinnedBlock) {} where the pin belonged, so its four tests have been red since 2026-09-01 , unnoticed because this branch has never had a PR for CI to run on.--deny-warnings failures, all older than this branch's tip.| Job | Result |
|---|---|
| autofix (brand, oxlint@1.78.0, deps:check) | pass |
Typecheck (turbo run types) | pass |
| Mail app tests (Jest) | 4,814 pass |
| Mail tests, non-UTC | 5 pass |
| Server tests (Vitest) | 12,639 pass |
| Deploy scope (bash) | 16 pass |
staging. The branch is 324 commits behind and the merge conflicts in CompanyExplorer.tsx, FilesTab.tsx, TableDocumentView.tsx and TableGrid.tsx. It needs a quiet tree.chatHistoryList.test.ts's bucketing passes under UTC and fails under PDT, so the UTC runner structurally cannot catch it. Not fixed here.🤖 Generated with Claude Code
This PR is not safe to merge until the agent-invitation authorization bypass, cross-team unfiling, unvalidated agent filing, and onboarding error redirect are fixed.
| Surface | Requests | Errors | p95 | Users | Verdict |
|---|---|---|---|---|---|
| db.select | 1654 → 6583 | 3 → 1 (0.0%) | 1098 ms → 50 ms | 0 | Working 6,583 requests since the deploy with 1 errors (3 errors in the 1,654-request baseline before it). p95 50 ms, was 1098 ms. |
| db.select | 1891 → 2371 | 0 → 0 (0%) | 83 ms → 42 ms | 0 | Working 2,371 requests since the deploy with 0 errors (0 errors in the 1,891-request baseline before it). p95 42 ms, was 83 ms. |
| model_chunk Conversation Field Updater | 0 → 2 | 0 → 0 (0%) | not measured → 1 ms | 0 | Insufficient traffic 2 requests, under the 20 Libra needs |
| execute_tool update-conversation-fields | 0 → 2 | 0 → 0 (0%) | not measured → 1210 ms | 0 | Insufficient traffic 2 requests, under the 20 Libra needs |
| cron-slow-task-queue.pollCedarInputCrawlers | 0 → 1 | 0 → 0 (0%) | not measured → 79 ms | 0 | Insufficient traffic 1 request, under the 20 Libra needs |
| cron-slow-task-queue.pollCedarInputCrawlers | 0 → 1 | 0 → 0 (0%) | not measured → 287 ms | 0 | Insufficient traffic 1 request, under the 20 Libra needs |
| model_step Conversation Field Updater | 0 → 0 | 0 → 0 (0%) | not measured | 0 | No traffic No requests recorded since this deploy. |
| /api/trpc/crm.getConversationFieldDefinitions | 0 → 0 | 0 → 0 (0%) | not measured | 0 | No traffic No requests recorded since this deploy. |
| model_chunk Conversation Field Updater | 4 → 0 | 0 → 0 (0%) | 1 ms → not measured | 0 | No traffic No requests recorded since this deploy. |
| execute_tool update-conversation-fields | 1 → 0 | 0 → 0 (0%) | 849 ms → not measured | 0 | No traffic No requests recorded since this deploy. |
| model_step Conversation Field Updater | 2 → 0 | 0 → 0 (0%) | 7325 ms → not measured | 0 | No traffic No requests recorded since this deploy. |
| /api/trpc/crm.getConversationFieldDefinitions | 3 → 0 | 0 → 0 (0%) | 355 ms → not measured | 0 | No traffic No requests recorded since this deploy. |