Skip to content

Compliance audit — 2026-08-07 #958

Closed

Description

@don-petry

Compliance Audit — 2026-08-07

This umbrella issue tracks all findings from the automated compliance audit run on 2026-08-07.
Findings are grouped by remediation category. Address each category together to avoid duplicate agent PRs.

Total findings: 65 across 12 repositories


Remediation Work Breakdown

Repository Settings (5 finding(s))

Remediation: apply-repo-settings.sh
Affected repos: demo-repository

Repo Check Severity
demo-repository allow_auto_merge warning
demo-repository delete_branch_on_merge warning
demo-repository has_discussions error
demo-repository missing-codeowners error
demo-repository check-suite-prefs-unreadable warning

Push Protection & Secret Scanning (6 finding(s))

Remediation: apply-repo-settings.sh (security_and_analysis) + per-repo ci.yml and .gitignore
Affected repos: .github-private, demo-repository, markets, repo-template

Repo Check Severity
repo-template gitignore_baseline error
.github-private gitignore_baseline error
markets gitignore_baseline error
demo-repository security_and_analysis_unavailable warning
demo-repository secret_scan_ci_job_present error
demo-repository gitignore_baseline error

Repository Rulesets (28 finding(s))

Remediation: apply-rulesets.sh
Affected repos: .github, .github-private, ContentTwin, TalkTerm, bmad-bgreat-suite, broodly, broodminder-export, demo-repository, google-app-scripts, incubator, markets, repo-template

Repo Check Severity
repo-template ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
repo-template ruleset-drift-pr-quality-require_last_push_approval error
.github ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
.github ruleset-drift-pr-quality-require_last_push_approval error
.github-private ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
.github-private ruleset-drift-pr-quality-require_last_push_approval error
markets ruleset-bypass-orgadmin-code-quality error
markets ruleset-bypass-dependabot-code-quality error
markets ruleset-drift-pr-quality-require_code_owner_review error
markets ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
markets ruleset-drift-pr-quality-require_last_push_approval error
broodly ruleset-drift-pr-quality-require_code_owner_review error
broodly ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
broodly ruleset-drift-pr-quality-require_last_push_approval error
google-app-scripts ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
google-app-scripts ruleset-drift-pr-quality-require_last_push_approval error
ContentTwin ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
ContentTwin ruleset-drift-pr-quality-require_last_push_approval error
incubator ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
incubator ruleset-drift-pr-quality-require_last_push_approval error
TalkTerm ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
TalkTerm ruleset-drift-pr-quality-require_last_push_approval error
bmad-bgreat-suite ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
bmad-bgreat-suite ruleset-drift-pr-quality-require_last_push_approval error
demo-repository missing-pr-quality error
demo-repository missing-code-quality error
broodminder-export ruleset-drift-pr-quality-dismiss_stale_reviews_on_push error
broodminder-export ruleset-drift-pr-quality-require_last_push_approval error

Workflows (18 finding(s))

Remediation: per-repo workflow additions
Affected repos: .github-private, ContentTwin, demo-repository, google-app-scripts, markets

Repo Check Severity
.github-private non-stub-agent-shield.yml error
markets non-stub-feature-ideation.yml error
google-app-scripts stub-surface-drift-pr-auto-review.yml-on error
ContentTwin stub-surface-drift-dependency-audit.yml-concurrency warning
ContentTwin stub-surface-drift-pr-review-mention.yml-concurrency warning
ContentTwin stub-surface-drift-pr-auto-review.yml-concurrency warning
demo-repository missing-ci.yml error
demo-repository missing-sonarcloud.yml error
demo-repository missing-dev-lead.yml error
demo-repository missing-dependabot-automerge.yml error
demo-repository missing-dependency-audit.yml error
demo-repository missing-agent-shield.yml error
demo-repository missing-pr-review-mention.yml error
demo-repository missing-feature-ideation.yml error
demo-repository missing-pr-auto-review.yml error
demo-repository missing-initiative-driver.yml error
demo-repository missing-permissions-auto-assign.yml warning
demo-repository missing-permissions-proof-html.yml warning

Action SHA Pinning (2 finding(s))

Remediation: pin actions to SHA in each workflow file
Affected repos: demo-repository

Repo Check Severity
demo-repository unpinned-actions-auto-assign.yml error
demo-repository unpinned-actions-proof-html.yml error

Dependabot Configuration (2 finding(s))

Remediation: per-repo .github/dependabot.yml
Affected repos: demo-repository, repo-template

Repo Check Severity
repo-template missing-github-actions-ecosystem error
demo-repository missing-config error

Agent Standards (CLAUDE.md / AGENTS.md / copilot-setup-steps.yml) (4 finding(s))

Remediation: per-repo doc and workflow additions
Affected repos: demo-repository

Repo Check Severity
demo-repository missing-claude-md error
demo-repository missing-agents-md error
demo-repository missing-copilot-setup-steps warning
demo-repository missing-copilot-instructions error

Generated by the weekly compliance audit on 2026-08-07 13:28 UTC.
Address each remediation category as a single coordinated PR to avoid duplicate agent work.

Activity

  1. added
    compliance-auditAutomated compliance audit finding
    dev-leadFor dev-lead agent pickup
    and removed
    dev-leadFor dev-lead agent pickup
    on Aug 7, 2026
  2. don-petry commented on Aug 12, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead: Scope Analysis — no in-repo code change applies

    I picked this issue up as a code-implementing dev-lead and worked through Phase 1 (scope). Reporting the finding before touching any files, because none of the 65 findings map to a source change in petry-projects/.github itself.

    Why

    Category Findings Remediation In-repo file change?
    Repository Rulesets 28 apply-rulesets.sh (admin token) No — codified standards/rulesets/pr-quality.json already carries both drifted keys (dismiss_stale_reviews_on_push: true, require_last_push_approval: true). The drift is live-side only.
    Repository Settings 5 apply-repo-settings.sh (admin API) No — live settings on demo-repository
    Push Protection & Secret Scanning 6 apply-repo-settings.sh + per-repo ci.yml/.gitignore No — file additions to repo-template, .github-private, markets, demo-repository
    Workflows 18 deploy standard workflows No — additions to other repos (mostly demo-repository); source templates already exist in standards/workflows/
    Action SHA Pinning 2 pin actions in workflow files No — demo-repository files
    Dependabot 2 per-repo .github/dependabot.yml No — demo-repository, repo-template; templates exist in standards/dependabot/
    Agent Standards 4 per-repo docs No — demo-repository docs

    Every remediation is either a live admin-API operation or a file addition to a different repository. The audit (scripts/compliance-audit.sh:735) treats the codified standards/rulesets/*.json as the source of truth, and it is already correct. standards/ruleset-remediation-runbook.md explicitly notes ruleset changes are not auto-remediated — they require an admin token the CI GITHUB_TOKEN does not carry.

    Correct remediation (operator with admin token)

    export GH_TOKEN="$(gh auth token)"   # admin:org + repo
    ORG=petry-projects
    
    # Ruleset drift (28) — converge live rulesets to the codified standard
    for r in .github .github-private ContentTwin TalkTerm bmad-bgreat-suite broodly \
             broodminder-export google-app-scripts incubator markets repo-template demo-repository; do
      ./scripts/apply-rulesets.sh --repo "$ORG/$r"
    done
    ./scripts/fix-ruleset-bypass.sh markets   # bypass-actor findings on code-quality
    
    # Repo settings + push protection (11)
    ./scripts/apply-repo-settings.sh --repo "$ORG/demo-repository"
    
    # demo-repository is missing the full standard surface (workflows, dependabot,
    # CLAUDE.md/AGENTS.md/copilot, CODEOWNERS) — bootstrap it:
    ./scripts/bootstrap-new-repo.sh demo-repository
    ./scripts/sync-gitignore-baseline.sh repo-template .github-private markets demo-repository

    Recommendation

    This umbrella was routed to a code-implementing dev-lead, but it is an operational, multi-repo task. There is no file in this repo I can edit that closes any finding — the standards that source the remediations are already present and correct. I am not fabricating an in-repo PR (e.g., copying demo-repository's files into this standards repo), as that would pollute standards/ without remediating anything live.

    Closing my run with no file changes. Suggest routing to an operator/admin-token workflow, or splitting per-repo file additions (demo-repository, repo-template) into their own repo-scoped issues.

  3. added
    dev-leadFor dev-lead agent pickup
    and removed
    dev-leadFor dev-lead agent pickup
    on Aug 12, 2026
  4. don-petry commented on Aug 12, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead Implementation Plan

    Issue: #958 — Compliance audit — 2026-08-07

    Scope

    This PR closes the only change actionable inside petry-projects/.github itself:
    lock the codified pr-quality ruleset's dismiss_stale_reviews_on_push parameter
    with an anti-divergence guard test — the direct companion to #957, which locked
    require_last_push_approval.

    Why the other 63 findings are out of scope for this PR:

    • Per the repo-boundary rule (docs(governance): codify repo boundary — org-wide rulesets belong in .github, not .github-private (#575) #576, see standards/github-settings.md), .github
      owns policy; it cannot edit files in other repos (markets, demo-repository,
      broodly, …). Those per-repo workflow/settings/dependabot findings are fixed by
      PRs in each target repo, not here.
    • The ruleset-drift-* findings are live drift between each repo's GitHub-side
      ruleset and the codified source of truth. The canonical
      standards/rulesets/pr-quality.json already sets dismiss_stale_reviews_on_push: true
      and require_last_push_approval: true. Converging the live rulesets is a manual,
      admin-token operation
      (apply-rulesets.sh, per the Ruleset Remediation Runbook) —
      not a code change, and explicitly skipped by compliance-remediate.sh.
    • The in-repo code deliverable for a ruleset-drift-pr-quality-<param> finding is the
      guard test that prevents the codified source of truth from silently regressing, so
      the applier can never push a drifted value fleet-wide. Compliance: ruleset-drift-pr-quality-require_last_push_approval #957 established this pattern.

    Implementation Checklist

    Tests to Write

    • New bats guard: pr-quality.json: dismisses stale reviews on push (dismiss_stale_reviews_on_push == true)
      — proven red by temporarily flipping the codified value to false, then green once restored

    Test Command

    ./node_modules/.bin/bats test/scripts/apply-rulesets/apply-rulesets.bats

  5. don-petry commented on Aug 12, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead: Implementation Complete

    Plan Execution

    • Added dismiss_stale_reviews_on_push == true anti-divergence guard to test/scripts/apply-rulesets/apply-rulesets.bats, mirroring the Compliance: ruleset-drift-pr-quality-require_last_push_approval #957 require_last_push_approval guard
    • Proved the guard red (flipped codified value to false → test failed), then restored the source of truth
    • Full bats suite + required dev-lead-lint.sh green

    Scope note

    This is the only change actionable inside petry-projects/.github for issue #958.
    The .github repo owns policy but cannot edit other repos (#576), and ruleset-drift-*
    findings are live drift converged by the manual apply-rulesets.sh admin procedure —
    not a code change. standards/rulesets/pr-quality.json already codifies
    dismiss_stale_reviews_on_push: true; the in-repo deliverable is the guard that stops the
    codified source of truth from silently regressing (same pattern #957 set for
    require_last_push_approval). The remaining 63 umbrella findings belong to per-repo PRs.

    Test Results

    # target suite (was 22 tests, now 23)
    $ ./node_modules/.bin/bats test/scripts/apply-rulesets/apply-rulesets.bats
    ok 5 pr-quality.json: requires last-push approval (require_last_push_approval == true)
    ok 6 pr-quality.json: dismisses stale reviews on push (dismiss_stale_reviews_on_push == true)
    ...
    23 tests, 0 failures
    
    # red proof (codified value temporarily flipped to false)
    not ok 1 pr-quality.json: dismisses stale reviews on push (dismiss_stale_reviews_on_push == true)
    #   `[ "$output" = "true" ]' failed   → source of truth restored afterward
    
    # full repo bats suite
    1134 tests, 0 failures  (exit 0)
    
    # required lint
    $ bash .dev-lead/scripts/dev-lead-lint.sh
    [lint] shellcheck on 29 file(s)... OK
    [lint] all checks passed
    

    Files Changed

    • test/scripts/apply-rulesets/apply-rulesets.bats: added one anti-divergence guard test asserting the codified pr-quality ruleset keeps dismiss_stale_reviews_on_push == true

    Notes

    Phase 2 TDD note: the codified JSON already satisfies the invariant, so "red" was
    demonstrated by temporarily flipping the value to false and confirming the new guard
    fails, then restoring it. No production/source files were modified — the change is a
    regression lock only.

  6. don-petry commented on Aug 12, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead: cannot implement issue #958 — needs human attention

    A stage timed out (exit 124) while implementing this issue. A timeout is treated as a first-class, non-retryable condition — retrying with the same budget would just time out again — so this escalates for human review on the first occurrence instead of wasting a same-budget retry.

    This hit the max action budget. Split this issue into smaller, independently-implementable pieces, or raise vars.ACTION_TIMEOUT_SEC, then re-apply the dev-lead label.

    Note: the engine had produced changes before the timeout. They were on branch dev-lead/issue-958-20260812-0539 on the runner and were not pushed, so they are not recoverable from this run (the runner is ephemeral). The redacted session snippet below shows what it attempted — treat it as context, not restorable output. Re-apply dev-lead after splitting the issue or raising the budget to regenerate the work.

  7. added
    dev-leadFor dev-lead agent pickup
    and removed
    dev-leadFor dev-lead agent pickup
    on Aug 12, 2026
  8. don-petry commented on Aug 12, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead Implementation Plan

    Issue: #958 — Compliance audit — 2026-08-07

    Scope

    Lock the codified pr-quality ruleset's dismiss_stale_reviews_on_push parameter with an
    anti-divergence guard test — the direct companion to #957, which locked
    require_last_push_approval. This is the only change actionable inside
    petry-projects/.github itself for this umbrella.

    Why the other 63 findings are out of scope for this PR:

    • Repo-boundary rule (docs(governance): codify repo boundary — org-wide rulesets belong in .github, not .github-private (#575) #576, standards/github-settings.md): .github owns policy; it
      cannot edit files in other repos (markets, demo-repository, broodly, …). Per-repo
      workflow/settings/dependabot/doc findings are fixed by PRs in each target repo.
    • ruleset-drift-* findings are live drift between each repo's GitHub-side ruleset and
      the codified source of truth. standards/rulesets/pr-quality.json already sets
      dismiss_stale_reviews_on_push: true (and require_last_push_approval: true).
      Converging live rulesets is a manual, admin-token operation (apply-rulesets.sh,
      per the Ruleset Remediation Runbook) — not a code change, and explicitly skipped by
      compliance-remediate.sh.
    • The in-repo code deliverable for a ruleset-drift-pr-quality-<param> finding is the guard
      test that prevents the codified source of truth from silently regressing. Compliance: ruleset-drift-pr-quality-require_last_push_approval #957 set this
      pattern; this PR extends it to the sibling parameter.

    Implementation Checklist

    Tests to Write

    • New bats guard: pr-quality.json: dismisses stale reviews on push (dismiss_stale_reviews_on_push == true) — proven red by temporarily flipping the
      codified value to false, then green once restored

    Test Command

    ./node_modules/.bin/bats test/scripts/apply-rulesets/apply-rulesets.bats

  9. don-petry commented on Aug 12, 2026

    @don-petry
    ContributorAuthor

    Dev-Lead: Implementation Complete

    Plan Execution

    • Added dismiss_stale_reviews_on_push == true anti-divergence guard to test/scripts/apply-rulesets/apply-rulesets.bats, mirroring the Compliance: ruleset-drift-pr-quality-require_last_push_approval #957 require_last_push_approval guard
    • Proved the guard red (flipped codified value to false → test failed at [ "$output" = "true" ]), then restored the source of truth
    • Target suite, full bats suite, and required dev-lead-lint.sh all green

    Scope note

    This is the only change actionable inside petry-projects/.github for issue #958.
    The .github repo owns policy but cannot edit other repos (#576), and ruleset-drift-*
    findings are live drift converged by the manual apply-rulesets.sh admin procedure —
    not a code change. standards/rulesets/pr-quality.json already codifies
    dismiss_stale_reviews_on_push: true; the in-repo deliverable is the guard that stops the
    codified source of truth from silently regressing (same pattern #957 set for
    require_last_push_approval). The remaining per-repo findings (settings, workflows,
    dependabot, docs, SHA pinning, push-protection) belong to PRs in each target repo.

    Test Results

    # red proof (codified value temporarily flipped to false)
    $ ./node_modules/.bin/bats -f "dismisses stale reviews on push" test/scripts/apply-rulesets/apply-rulesets.bats
    not ok 1 pr-quality.json: dismisses stale reviews on push (dismiss_stale_reviews_on_push == true)
    #   `[ "$output" = "true" ]' failed   → source of truth restored afterward
    
    # target suite (now 23 tests, new guard is #6)
    $ ./node_modules/.bin/bats test/scripts/apply-rulesets/apply-rulesets.bats
    ok 5 pr-quality.json: requires last-push approval (require_last_push_approval == true)
    ok 6 pr-quality.json: dismisses stale reviews on push (dismiss_stale_reviews_on_push == true)
    ...
    23 tests, 0 failures
    
    # full repo bats suite
    ok count: 1134
    not-ok count: 0
    
    # required lint
    $ bash .dev-lead/scripts/dev-lead-lint.sh
      [lint] shellcheck on 29 file(s)...
      [lint] validating .github/CODEOWNERS...
      [lint] all checks passed
    

    Files Changed

    • test/scripts/apply-rulesets/apply-rulesets.bats: added one anti-divergence guard test asserting the codified pr-quality ruleset keeps dismiss_stale_reviews_on_push == true

    Notes

    Phase 2 TDD note: the codified JSON already satisfies the invariant, so "red" was
    demonstrated by temporarily flipping the value to false and confirming the new guard
    fails, then restoring it. No production/source files were modified — the change is a
    regression lock only. Prior attempt on this umbrella timed out (2100s) before pushing;
    this run reproduces the minimal in-repo deliverable only, so it stays well within budget.

  10. don-petry commented on Aug 14, 2026

    @don-petry
    ContributorAuthor

    Resolved! This check is now passing as of 2026-08-14. Closing automatically.

  11. added 2 commits that reference this issue on Aug 18, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    compliance-auditAutomated compliance audit findingdev-leadFor dev-lead agent pickupdev-lead:needs-humandev-lead could not complete this issue; needs human attention

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions