xiduzo/wtf / skills/wtf.pr-review
wtf.pr-review
VerifiedThis skill should be used when a developer or tech lead wants to review a pull request for correctness, spec adherence, and code quality — for example "review PR
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 9.3 KB |
| evals/evals.json | config | 3.7 KB |
| evals/fixtures/missing-test-coverage.json | config | 6.4 KB |
| evals/fixtures/schema-drift.json | config | 4.6 KB |
| evals/fixtures/scope-creep.json | config | 6.6 KB |
evals/evals.json
View on GitHub{
"skill_name": "wtf.pr-review",
"evals": [
{
"id": 1,
"prompt": "Run wtf.pr-review in test mode. Skip steps 0–4 (gh setup, PR identification, spec hierarchy fetch, diff fetch). Use the fixture from `evals/fixtures/missing-test-coverage.json` — its `trace` field is the Trace spec, its `feature` field carries the canonical text of the claimed scenarios, and its `diff` and `diff_stat` fields replace the output of `git diff`. Run step 5 only (review checklist). Output all findings.",
"expected_output": "A structured review with a FAIL or WARN on the Spec adherence dimension, explicitly identifying that the second claimed scenario (automatic cancellation after the grace period) has no corresponding test. Verdict must be REQUEST CHANGES or APPROVE WITH COMMENTS. Scenario 1 (timeout warning) should be noted as covered.",
"files": ["evals/fixtures/missing-test-coverage.json"],
"expectations": [
"The Spec adherence dimension is rated FAIL or WARN",
"The missing test is identified by reference to cancellation or grace period (the second claimed scenario)",
"Scenario 1 (timeout warning) is noted as covered or passing",
"The verdict is REQUEST CHANGES or APPROVE WITH COMMENTS — not a clean APPROVE"
]
},
{
"id": 2,
"prompt": "Run wtf.pr-review in test mode. Skip steps 0–4 (gh setup, PR identification, spec hierarchy fetch, diff fetch). Use the fixture from `evals/fixtures/schema-drift.json` — its `trace` field is the Trace spec, its `feature` field carries the canonical text of the claimed scenarios, and its `diff` and `diff_stat` fields replace the output of `git diff`. Run step 5 only (review checklist). Output all findings.",
"expected_output": "A structured review with a FAIL on the Contract compliance dimension. The review must call out the field name mismatches (id vs settlementId, value vs amount, state vs status) and note that the currency field is missing from the implementation. Verdict must be REQUEST CHANGES.",
"files": ["evals/fixtures/schema-drift.json"],
"expectations": [
"The Contract compliance dimension is rated FAIL",
"The mismatch between `settlementId` (spec) and `id` (implementation) is explicitly named",
"The missing `currency` field is identified as a contract violation",
"The verdict is REQUEST CHANGES"
]
},
{
"id": 3,
"prompt": "Run wtf.pr-review in test mode. Skip steps 0–4 (gh setup, PR identification, spec hierarchy fetch, diff fetch). Use the fixture from `evals/fixtures/scope-creep.json` — its `trace` field is the Trace spec, its `feature` field carries the canonical scenario text and the Trace Plan, and its `diff` and `diff_stat` fields replace the output of `git diff`. Run step 5 only (review checklist). Output all findings.",
"expected_output": "A structured review with a WARN or FAIL on the Spec adherence or Spine Position dimension, specifically naming src/auth/middleware.ts as a change outside this Trace's Scenario Claim. The finding should note the token revocation logic implements the 'Revoked session cannot export settlements' scenario, which the Trace Plan assigns to a different Trace (#26). Verdict is at minimum APPROVE WITH COMMENTS.",
"files": ["evals/fixtures/scope-creep.json"],
"expectations": [
"The Spec adherence or Spine Position dimension is rated WARN or FAIL",
"`src/auth/middleware.ts` is specifically named as an unexpected changed file",
"The finding states the change implements a scenario outside the Trace's Scenario Claim, or that it belongs to another Trace in the Trace Plan",
"The verdict is APPROVE WITH COMMENTS or REQUEST CHANGES — not a clean APPROVE"
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |