version: "1.0.1" name: pr-reviewer description: >- Reviews the local diff or branch and returns a read-only, severity-tiered findings report. Modes cover standard bugs, structural quality, AI slop, and security audit. Use when asked to run /pr-reviewer, "review my changes", "code review", "thermo-nuclear review", "structural review", "deslop this", "clean up AI code", "security audit", "find vulnerabilities", or before commit, push, or handoff. For fixes use tidy; for PR creation use pr-creator; for CI or review comments use pr-babysitter; for frontend UX, accessibility, layout, state coverage, or rendered quality use ui-audit; for plans use planning.

Local Review

• IS: read-only review of a local diff, branch diff, or PR. Returns severity-tiered findings; leaves the working tree unchanged.
• IS NOT: fixing code (tidy), creating PRs (pr-creator), monitoring CI or review threads (pr-babysitter), frontend PR UX/accessibility/rendered-quality review (ui-audit), architecture briefs (define-architecture), reviewing plans (planning).

Only report issues you can defend with file:line evidence.

Mode dispatch

Pick one mode from the user's wording; load only its references:

Conditional loads:

• references/security-checklist.md for auth, input handling, external APIs, uploads, or environment config.
• references/performance-checklist.md for fetching, rendering, images, dependencies, or bundle-affecting imports.
• references/comment-examples.md for output-format calibration.
• agents/openai.yaml only for the optional second-opinion pass below.

Workflow

Review progress:
- [ ] Dispatch mode and load references
- [ ] Discover target
- [ ] Gather context and baseline checks
- [ ] Review, shard if needed
- [ ] Validate exact lines
- [ ] Report

1. Discover target. Review staged/unstaged changes first; if clean, review the branch diff. For a PR, use the same criteria and the PR handoff format.
2. Gather context. Capture intent. Load scoped AGENTS.md / CLAUDE.md. Run documented lint, type-check, and tests where they exist; record baseline failures.
3. Review. Apply the loaded rubric and high-signal criteria; shard large diffs. Optional: run an installed different-model CLI (codex exec, droid exec) read-only with agents/openai.yaml's default_prompt, then validate its findings.
4. Validate. Re-check exact lines. Drop speculative, duplicate, misread, mis-attributed, or pre-existing findings. Diff modes require changed lines; Security audit requires real in-scope code.
5. Report. Use references/severity-rubric.md; structural blockers go under Must fix before push.

High-signal criteria

Report only when certain:

• Compile, type, import, or syntax failure.
• Clear runtime bug, state error, or data-handling regression.
• Concrete exploit path, named vulnerability class, and affected file:line.
• Measurable performance regression.
• Missing necessary tests: render-only checks for interactive behavior, or bug fixes without a failing repro test at the seam that failed (route, API, integration point, not a helper invented during the fix).
• Test setup that requires helper tracing to understand assertions.
• New lint, type-check, or test failures versus baseline.
• Scoped instruction-file violation, with the rule quoted.
• Speculative abstraction or avoidable complexity without a current requirement.
• AI-generated patterns from references/ai-slop-patterns.md.
• File pushed past ~1000 lines when the new behavior has a local module, component, or helper boundary.
• Feature-specific conditionals added to unrelated shared paths.
• Bespoke helper duplicating a canonical utility.
• Hand-rolled reimplementation of stdlib or native platform behavior, with the replacement named.
• New dependency added for what the stdlib, the platform, or an already-installed dependency covers.
• Logic in the wrong layer when the canonical home is clear.
• Retried or at-least-once write with no idempotency key or dedupe barrier, so a duplicate delivery applies twice.
• Database commit plus an external publish (queue, webhook, email) without an outbox or transactional guarantee (dual-write): one side can fail independently.
• External input (webhook/callback) trusted blindly: signature not verified over the raw bytes, or state overwritten directly from the payload instead of confirming against the source.
• Floats or other lossy types used for money or precision-sensitive values, or money serialized as a bare JSON number rather than a string or integer minor-units.
• Multi-step flow with an irreversible external effect and no compensation or resume path if a later step fails.
• Sensitive mutation (funds, permissions, config) with no audit trail of what changed, who changed it, and why.

Do not report style preferences, unrelated pre-existing issues, risks without a repro or exploit path, broad rewrites outside the diff's intent, linter-only noise, or explicitly silenced violations.

Output

Default local report:

## Local review
 
### Must fix before push
-[<severity>] `path/to/file.ts:line` <short factual title>
  Why: <concrete impact>
  Fix: <committable fix>
 
### Should fix soon
-[<severity>] `path/to/file.ts:line` <short factual title>
  Why: <concrete impact>
  Fix: <committable fix>
 
### Ready for handoff
-<readiness summary, including lint/type-check/test baseline>

If no issues, write None. under the first two tiers and state what was checked.

PR handoff format:

## PR handoff summary
 
-[<severity>] `path/to/file.ts:line` <short factual title>
  Why: <concrete impact>
  Fix: <committable fix>

Gotchas

• Local changes beat PR lookup; review the working tree first or you miss uncommitted bugs.
• Loading every rubric makes standard review noisy. Load only the selected mode.
• Editing mid-review breaks the contract. Route fixes to tidy.
• Findings without file:line, concrete impact, and a fix do not belong in the report.
• Merge duplicate findings that share one root cause.
• External-engine output is advisory. Re-validate everything against the actual diff.
• Skipping the baseline makes pre-existing failures look like regressions.

Related skills

• tidy: applies fixes in place and verifies them.
• pr-creator: creates or updates the PR after review.
• pr-babysitter: monitors CI and inbound review comments.
• ui-audit: frontend PR review for user-facing UX, accessibility, layout, state coverage, and rendered quality.
• define-architecture: forward-looking architecture briefs and adoption opportunities outside a diff review.
• planning: builds and reviews plans before implementation.