Skip to content

Daily Autohealing Report — 2026-05-25 (UTC) #3375

@fro-bot

Description

@fro-bot

Daily Autohealing Report — 2026-05-25 (UTC)

Errored PRs

None. Open PR queue is empty. Yesterday's #3371 (pnpm overrides migration) merged at 2026-05-24T13:26Z.

Security

None requiring action. 0 open Dependabot alerts. Three OpenSSF Scorecard meta-findings still open (Fuzzing #7, CII-Best-Practices #5, Branch-Protection #1). Unchanged.

Health & Maintenance

  • Failing scheduled workflow: Merge Data Branch (run 26375057529 at 2026-05-24T22:54Z, re-dispatch 26390324135 at 2026-05-25T08:05Z failed identically). Root cause is not transient and not in the workflow — it's two private wiki entries in metadata/repos.yaml whose GraphQL node IDs return NOT_FOUND:

    • R_kgDOSVJgdw[REDACTED] (added 2026-05-05, onboarding_status: pending)
    • R_kgDOSZ9x-w[REDACTED] (added 2026-05-13, onboarding_status: pending)

    Either the repos were deleted/transferred before onboarding completed, or the App invitation was never accepted. Verified with a direct GraphQL probe — GitHub returns {"data":{"node":null},"errors":[{"type":"NOT_FOUND",...}]} for both.

    Diagnostic gap noted: scripts/check-wiki-private-presence.ts:144 labels both failures [subprocess-threw] — investigate network/rate-limit/auth, but the actual mode is node-null (repo lifecycle/App access). The bare catch at line 133 fires before the isGraphQLNodeNullResponse branch (line 127) can dispatch — gh api graphql exits non-zero on GraphQL errors, so node-null cases get misclassified as subprocess failures. Mentioning this for human awareness; not patching, since the fail-closed safety property is intentional and any change to the script's error semantics deserves Marcus's review.

Developer Experience

None. Repository validation green:

  • pnpm run lint — clean
  • pnpm run check-types — clean
  • pnpm test — 22 files / 644 passed + 3 todo in ~3s

Needs Human Attention

  • Merge Data Branch will keep failing until the two orphan entries are resolved. Choose one:

    1. Complete onboarding for the two pending repos (re-issue App invitations).
    2. Remove the orphaned entries from metadata/repos.yaml.

    The weekly sweep is currently blocked end-to-end.

  • Optional follow-up: improve the diagnostic in check-wiki-private-presence.ts so future operators see "investigate repo lifecycle/App access" instead of the wrong network hint. Small targeted change; could be a one-line PR if you want it.

  • OpenSSF Scorecard alerts (Fuzzing Update peter-murray/workflow-application-token-action to v2 #7, CII-Best-Practices feat: add Renovate workflow for dependency auto-updates #5, Branch-Protection feat: set default settings #1) — same three meta-recommendations across many cycles. Implement or dismiss.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions