Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Fetch unresolved PR review comments using GitHub GraphQL API, filtering out resolved and outdated feedback.
.claude/skills/microck-fetch-unresolved-comments/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 103% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -49% | 0% |
Uses GitHub's GraphQL API to fetch only unresolved review thread comments from a pull request.
PR_NUMBER and GITHUB_REPOSITORY are set, read them and parse GITHUB_REPOSITORY as owner/repo and use PR_NUMBER directlygh pr view --json url -q '.url' to get the current branch's PR URL and parse to extract owner, repo, and PR numberbash GITHUB_TOKEN=$(gh auth token) \ uv run python .claude/skills/fetch-unresolved-comments/fetch_unresolved_comments.py <owner> <repo> <pr_number>
--token <token>: Provide token explicitly (default: GITHUB_TOKEN or GH_TOKEN env var)The script always outputs JSON with:
total: Total number of unresolved comments across all threadsby_file: Review threads grouped by file path (each thread contains multiple comments in a conversation)json{ "total": 3, "by_file": { ".github/workflows/resolve.yml": [ { "thread_id": "PRRT_kwDOAL...", "isOutdated": false, "line": 40, "startLine": null, "diffHunk": "@@ -0,0 +1,245 @@\n+name: resolve...", "comments": [ { "id": 2437935275, "body": "We can remove this once we get the key.", "author": "harupy", "createdAt": "2025-10-17T00:53:20Z" }, { "id": 2437935276, "body": "Good catch, I'll update it.", "author": "contributor", "createdAt": "2025-10-17T01:10:15Z" } ] } ], ".gitignore": [ { "thread_id": "PRRT_kwDOAL...", "isOutdated": false, "line": 133, "startLine": null, "diffHunk": "@@ -130,0 +133,2 @@\n+.claude/*", "comments": [ { "id": 2437935280, "body": "Should we add this to .gitignore?", "author": "reviewer", "createdAt": "2025-10-17T01:15:42Z" } ] } ] } }
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | pass→pass | 7,378 | 2,919 | -60% | 1 | 1 | 0% | 1,352 | 1,267 | -6% | 0 | 0 | — |
case-01 | fail→fail | 10,254 | 5,456 | -47% | 1 | 1 | 0% | 1,933 | 1,088 | -44% | 0 | 0 | — |
case-02 | fail→fail | 7,762 | 4,740 | -39% | 1 | 1 | 0% | 1,441 | 992 | -31% | 0 | 0 | — |
case-03 | fail→fail | 4,781 | 5,377 | +12% | 1 | 1 | 0% | 624 | 1,017 | +63% | 0 | 0 | — |
case-04 | pass→fail | 8,861 | 5,113 | -42% | 1 | 1 | 0% | 1,752 | 1,016 | -42% | 0 | 0 | — |
case-05 | fail→pass | 11,251 | 3,912 | -65% | 1 | 1 | 0% | 2,057 | 1,521 | -26% | 0 | 0 | — |
case-06 | fail→pass | 6,697 | 2,625 | -61% | 1 | 1 | 0% | 1,107 | 1,181 | +7% | 0 | 0 | — |
case-08 | fail→pass | 3,638 | 2,594 | -29% | 1 | 1 | 0% | 581 | 1,177 | +103% | 0 | 0 | — |
case-09 | fail→pass | 4,883 | 1,530 | -69% | 1 | 1 | 0% | 732 | 954 | +30% | 0 | 0 | — |
case-10 | fail→pass | 11,810 | 1,288 | -89% | 1 | 1 | 0% | 1,813 | 932 | -49% | 0 | 0 | — |
case-11 | fail→pass | 9,367 | 1,814 | -81% | 1 | 1 | 0% | 1,519 | 1,019 | -33% | 0 | 0 | — |
case-12 | pass→fail | 8,551 | 1,184 | -86% | 1 | 1 | 0% | 1,283 | 878 | -32% | 0 | 0 | — |
case-13 | fail→pass | 6,762 | 1,756 | -74% | 1 | 1 | 0% | 996 | 964 | -3% | 0 | 0 | — |
case-14 | pass→pass | 4,863 | 1,476 | -70% | 1 | 1 | 0% | 718 | 950 | +32% | 0 | 0 | — |
case-15 | pass→pass | 11,286 | 3,695 | -67% | 1 | 1 | 0% | 1,944 | 1,338 | -31% | 0 | 0 | — |
case-16 | pass→pass | 4,207 | 2,327 | -45% | 1 | 1 | 0% | 694 | 1,164 | +68% | 0 | 0 | — |
case-17 | pass→pass | 8,611 | 3,045 | -65% | 1 | 1 | 0% | 1,392 | 1,254 | -10% | 0 | 0 | — |
case-18 | fail→pass | 8,299 | 2,211 | -73% | 1 | 1 | 0% | 1,264 | 1,068 | -16% | 0 | 0 | — |
case-19 | pass→pass | 4,702 | 1,311 | -72% | 1 | 1 | 0% | 700 | 903 | +29% | 0 | 0 | — |
case-20 | pass→pass | 8,072 | 3,197 | -60% | 1 | 1 | 0% | 1,452 | 1,265 | -13% | 0 | 0 | — |
case-21 | pass→pass | 8,824 | 4,727 | -46% | 1 | 1 | 0% | 1,623 | 1,482 | -9% | 0 | 0 | — |
case-22 | pass→pass | 10,891 | 9,192 | -16% | 1 | 1 | 0% | 2,033 | 1,775 | -13% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +27 percentage points is the difference between those two pass rates over the 18 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.