v0.6.101: 11 new knowledgebase connectors, slack scopes update, login refinements#4858
Conversation
waleedlatif1
commented
Jun 3, 2026
- fix(schedules): count usage lim error schedule as failed run (fix(schedules): count usage lim error schedule as failed run #4853)
- feat(connectors): add 11 knowledge base connectors (feat(connectors): add 11 knowledge base connectors #4849)
- fix(auth): show "account already exists" on duplicate email signup (fix(auth): show "account already exists" on duplicate email signup #4855)
- fix(slack): request reactions:read in OAuth URL, drop im:history (fix(slack): request reactions:read in OAuth URL, drop im:history #4856)
* fix(schedules): count usage lim error schedule as failed run * remove backoff logic
* feat(connectors): add 11 knowledge base connectors (Gong, Grain, Fathom, Granola, incident.io, Rootly, Ashby, Greenhouse, DocuSign, Monday, GitLab) * fix(granola): restore correct calendar_event field names (event_title/organiser/calendar_event_id/scheduled_start_time/scheduled_end_time/invitees) * fix(connectors): correct Grain id field/include flags and incident.io timestamp value shape (verified vs raw API specs) * feat(connectors): add scoping filters for Gong (host users), Fathom (meeting type/domain), Granola (folder/created-after) * feat(connectors): add verified scoping filters (Grain, Rootly, Ashby, Greenhouse, DocuSign, GitLab) + clarify Monday board scope * fix(connectors): docusign sandbox source URL, greenhouse scorecard cap, gong overlap; add incident.io status/mode filters * chore(connectors): remove non-TSDoc inline comments * fix(fathom): apply 14-day overlap to incremental created_after so late transcripts are recaptured * fix(fathom): cache sourceUrl in header so getDocument preserves it; return concrete contentHash (drop omit/cast) * fix(connectors): grain check transcript before formatting; gong cache date window across cursor pages * fix(connectors): fathom skip meetings with no transcript/summary; docusign cache from_date across cursor pages * fix(fathom): return metadata from getDocument (carried via header cache) for tag consistency * fix(connectors): grain cache filter window across pages, greenhouse guard NaN page, rootly skip empty incidents * feat(connectors): final audit — add verified scoping filters (grain team/type, granola+greenhouse createdBefore, gitlab milestone, rootly service/team/env), fix incident.io paused category + rootly severity-slug copy
…4855) * fix(auth): show "account already exists" on duplicate email signup * fix(auth): use exact path match for duplicate-email signup check
* fix(slack): request reactions:read in OAuth URL, drop im:history * chore(slack): update read-messages missing-scope message to drop im:history
|
The latest updates on your projects. Learn more about Vercel for GitHub. |
PR SummaryMedium Risk Overview Scheduled workflows: introduces shared Granola: workflow block and docs gain optional Slack read-messages: clarifies the missing-scope error—channel history scopes vs. no Sim-bot DM history. Reviewed by Cursor Bugbot for commit b58cd1f. Configure here. |
Greptile SummaryThis PR bundles four separate improvements: 11 new knowledge-base connectors (Ashby, DocuSign, Fathom, GitLab, Gong, Grain, Granola, Greenhouse, incident.io, Monday, Rootly), Slack OAuth scope updates (
Confidence Score: 4/5Safe to merge with minor caveats; no blocking issues found. The schedule fix and connector additions are well-structured and the core logic is sound. The auth middleware email check has a subtle case-sensitivity gap where the lowercased input is compared against potentially mixed-case stored emails via a case-sensitive
Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[Schedule tick claims due schedule] --> B[preprocessExecution]
B --> C{HTTP status code}
C -- 401 --> D[Disable schedule]
C -- 403 --> D
C -- 404 --> D
C -- 429 --> E[Retry in 5 min, reset infraRetry]
C -- 402 Usage limit --> F[buildScheduleFailureUpdate\nnextRunAt = cron next OR now+1h]
C -- 5xx retryable --> G[retryScheduleAfterInfraFailure\nbackoff + jitter]
C -- other 5xx --> H[buildScheduleFailureUpdate\nnextRunAt = deployment or now+24h]
C -- success --> I[runWorkflowExecution]
F --> J[Increment failedCount\nSet lastFailedAt\nAuto-disable if >= MAX_CONSECUTIVE_FAILURES]
H --> J
I --> K{Execution result}
K -- success --> L[Reset failedCount, advance nextRunAt]
K -- failure --> M[buildScheduleFailureUpdate\nnextRunAt from blocks]
M --> J
G --> N{Retry attempts exhausted?}
N -- yes --> H
N -- no --> O[Set nextRunAt = now + backoff\nIncrement infraRetryCount]
Reviews (1): Last reviewed commit: "fix(slack): request reactions:read in OA..." | Re-trigger Greptile |
… emitting a degraded, un-refreshable record (#4859)