Skip to content

Turn any idea, plan, or codebase into a beautiful interactive diagram. An agent skill for Claude Code, Codex, and more.

10.5 KB
Raw Blame
# yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json
# Advisory pilot: retain human review and existing CI/branch protection.
language: en-US
tone_instructions: >-
  Be concise and constructive. Prioritize concrete defects and material evidence gaps.
  Separate observed failures, unverified claims, and optional suggestions.
  Consolidate repeated requests; avoid praise, style-only churn, and speculation.
reviews:
  profile: chill
  request_changes_workflow: false
  high_level_summary: true
  high_level_summary_in_walkthrough: true
  changed_files_summary: false
  high_level_summary_instructions: >-
    In at most 100 words, summarize the user problem, changed behavior, and
    compatibility impact without a file-by-file inventory. State the reviewed
    base/head and distinguish author-reported validation, reused evidence at its
    original revision, and observed CI at its checked revision. Treat the summary
    as a snapshot, not a live CI status. Do not claim browser or perceptual
    acceptance from static tests.
  poem: false
  in_progress_fortune: false
  sequence_diagrams: false
  assess_linked_issues: false
  suggested_labels: false
  suggested_reviewers: false
  auto_review:
    base_branches: ["^dev$"]
    enabled: true
    drafts: false
    auto_incremental_review: true
    auto_pause_after_reviewed_commits: 0
  finishing_touches:
    docstrings:
      enabled: false
    unit_tests:
      enabled: false
    autofix:
      enabled: false
  slop_detection:
    enabled: false
  # Exact generated examples checked by test/golden.mjs in the generated CI gate.
  # Filters also affect sparse checkout: keep runtime resources and authored inputs.
  # New outputs remain reviewable until their generation/freshness proof is checked.
  path_filters:
    - "!archify/examples/class-payment-processors-rendered.html"
    - "!archify/examples/class-payments-rendered.html"
    - "!archify/examples/dataflow-product-analytics.html"
    - "!archify/examples/erd-orders-rendered.html"
    - "!archify/examples/lifecycle-agent-run.html"
    - "!archify/examples/sequence-cache-miss-request.html"
    - "!archify/examples/subscription-billing-rendered.html"
    - "!archify/examples/timeline-archify-dev-activity-rendered.html"
    - "!archify/examples/timeline-payment-incident-rendered.html"
    - "!archify/examples/tree-archify-repository-rendered.html"
    - "!archify/examples/tree-payment-platform-rendered.html"
    - "!archify/examples/waterfall-checkout-request-rendered.html"
    - "!archify/examples/waterfall-example-rebuild-rendered.html"
    - "!archify/examples/web-app-rendered.html"
    - "!archify/examples/workflow-agent-tool-call-rendered.html"
    - "!examples/class-payment-processors-rendered.html"
    - "!examples/class-payments-rendered.html"
    - "!examples/dataflow-product-analytics.html"
    - "!examples/erd-orders-rendered.html"
    - "!examples/lifecycle-agent-run.html"
    - "!examples/sequence-cache-miss-request.html"
    - "!examples/subscription-billing-rendered.html"
    - "!examples/timeline-archify-dev-activity-rendered.html"
    - "!examples/timeline-payment-incident-rendered.html"
    - "!examples/tree-archify-repository-rendered.html"
    - "!examples/tree-payment-platform-rendered.html"
    - "!examples/waterfall-checkout-request-rendered.html"
    - "!examples/waterfall-example-rebuild-rendered.html"
    - "!examples/web-app-rendered.html"
    - "!examples/workflow-agent-tool-call-rendered.html"
  path_instructions:
    - path: "**"
      instructions: >-
        Use .agents/skills/archify-review/SKILL.md for value, cost, and impact
        judgments, and consult its linked repository guidance as needed. For changed
        tests or testing decisions, use .agents/skills/test-value/SKILL.md in scoped
        review mode and read its references when relevant. Identify the concrete
        mistake a proposed test catches, existing protection, and whether an
        equivalent valid result would fail. Recommend added coverage for a
        specific uncovered risk at the cheapest effective layer; account for
        surviving protection when tests are consolidated or removed. Apply
        the base branch's rules when evaluating proposed policy changes.
        Review the base-to-head source diff before generated output. The exact
        golden HTML copies listed in path_filters are excluded from review and
        may be absent from the sparse checkout. Use final-revision generated CI
        evidence for their freshness; request the specific missing evidence if
        unavailable. Source review and freshness checks do not establish visual
        acceptance. For each
        actionable finding, identify the trigger or material evidence gap, its
        consequence, and the smallest useful correction. Distinguish defects,
        unresolved claims, and optional suggestions. Accept equivalent prose
        and reachable documentation references; request another link or template
        entry only when it resolves a concrete ambiguity or missing requirement.
        On revision, read the latest PR body, discussion, and relevant CI. Focus
        on new changes and unresolved findings, reusing unaffected evidence with
        its original revision. Reuse maintainer scope decisions and consolidate
        requests already answered rather than repeating them inline.
    - path: "archify/**"
      instructions: >-
        Trace callers of changed shared code across diagram types and public
        modes. Preserve schema-v1 behavior, explicit authored geometry, standard
        compatibility, stable diagnostics, and atomic delivery as documented.
        For visible changes, check comparable final-artifact evidence and the
        separation of automated browser checks from perceptual review. For Viewer
        extraction, inspect initialization, shared state, cleanup, exports, and
        standalone HTML delivery. Do not propose a broad rewrite for a local fix.
    - path: ".github/workflows/**"
      instructions: >-
        Trace changed test dispatch through scope classification, shards,
        aggregation, and deployment or release consumers. Check that required
        coverage is complete, overlap is justified, and failed, cancelled or
        missing required work cannot become a successful aggregate. Accept
        intentional skips when the scope makes that work unnecessary. Distinguish
        runtime compatibility from native platform protection; inventory or mock
        checks prove dispatch, not real browser, OS or packaged execution.
        Evaluate speed claims using comparable measurements, including serial
        setup, uneven shard costs and resource contention; distinguish critical
        path from total compute. Recommend extra caching or orchestration only
        for an evidenced bottleneck, with proportionate maintenance cost.
    - path: "integrations/**"
      instructions: >-
        Check the package and release contracts in CONTRIBUTING.md. Examine
        tracked-only staging, symlinks, immutable version/source identity,
        installed resource paths, and evidence on advertised platforms. Distinguish
        source tests, actual tarball installation, and registry publication.
  pre_merge_checks:
    docstrings:
      mode: "off"
    title:
      mode: "off"
    description:
      mode: "off"
    issue_assessment:
      mode: "off" # Contribution scope below owns scope/issue assessment.
    custom_checks:
      - name: Contribution scope
        mode: warning
        instructions: >-
          Read CONTRIBUTING.md, REVIEWING.md, and .github/PULL_REQUEST_TEMPLATE.md. Pass when
          the PR explains the user problem, one focused behavior or delivery
          slice, compatibility/migration impact, and failure/rollback behavior
          where applicable, and the implementation fits that scope. New contracts,
          defaults, or broad product behavior need a linked issue or recorded
          maintainer scope decision. Narrow fixes may use a concrete reproduction;
          small documentation or test corrections may use a rationale. Do not
          require a PR to close every item of a tracking issue. Accept
          equivalent prose and omission of inapplicable template sections, not just
          exact headings or checked boxes. Warn once with the specific missing facts;
          read the latest description and maintainer decisions before repeating
          a warning. Do not classify missing evidence as a confirmed runtime defect.
      - name: Validation evidence
        mode: warning
        instructions: >-
          Assess evidence coverage and claim accuracy using the latest PR body,
          discussion, source diff, and relevant check results. Use CONTRIBUTING.md
          to identify the affected behavior and proportionate evidence. Pass when
          the available evidence covers the affected behavior and supports the
          material claims; warn for concrete coverage gaps or unsupported success
          claims. State the evaluated head
          and evidence revisions, distinguishing author reports, observed results,
          and reused evidence. Reuse unaffected results with a rationale; request
          reruns only where changes invalidate them. Accept equivalent prose,
          inapplicable sections being omitted, and linked evidence.
          Assess relevant regressions and generated-output freshness when inputs
          changed. For visible behavior, distinguish comparable artifact evidence,
          automated/browser checks, and perceptual review under the delivery contract.
          This check does not certify merge readiness: GitHub required checks and
          branch protection enforce final-head CI. Pending CI alone is not a
          coverage gap when other evidence is sufficient. If a pending or unavailable
          run is the only evidence for a material claim, identify that specific
          unresolved claim; never treat pending, skipped, or failed checks as passes.
          Report remaining gaps once with their consequence, smallest check, and
          owner. Fork-workflow approval belongs to maintainers. If repository or
          evidence access prevents assessment, return Inconclusive and describe
          the access limitation rather than claiming the author omitted evidence.
knowledge_base:
  code_guidelines:
    enabled: true
    filePatterns:
      - .agents/skills/archify-review/SKILL.md
      - .agents/skills/test-value/SKILL.md
      - CONTRIBUTING.md
      - REVIEWING.md
      - .github/PULL_REQUEST_TEMPLATE.md
      - archify/references/delivery-contract.md
  learnings:
    scope: local
  issues:
    scope: local
  pull_requests:
    scope: local