Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Draft changelog PRs by collecting GitHub/Slack/Git changes, formatting with templates, running quality gates, and preparing a branch/PR. Use when generating weekly/monthly release notes or when the user asks to create a changelog from recent merges. Trigger with "changelog weekly", "generate release notes", "draft changelog", "create changelog PR".
.claude/skills/jeremylongshore-changelog-orchestrator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -18% | 0% |
This skill turns raw repo activity (merged PRs, issues, commits, optional Slack updates) into a publishable changelog draft and prepares a branch/PR for review.
.changelog-config.json in the target repo.GITHUB_TOKEN for GitHub source).gh optional (used for PR creation if configured)..changelog-config.json from the repo root.${CLAUDE_SKILL_DIR}/scripts/validate_config.py.${CLAUDE_SKILL_DIR}/assets/weekly-template.md).${CLAUDE_SKILL_DIR}/scripts/render_template.py.date (ISO) and version (SemVer if known; otherwise 0.0.0).${CLAUDE_SKILL_DIR}/scripts/quality_score.py.output_path.changelog-YYYY-MM-DD, commit with docs: add changelog for YYYY-MM-DD.gh is configured, open a PR; otherwise, print the exact commands the user should run.See ${CLAUDE_SKILL_DIR}/references/implementation.md for detailed implementation guide.
CHANGELOG.md), plus an optional PR URL.${CLAUDE_SKILL_DIR}/scripts/quality_score.py.See ${CLAUDE_SKILL_DIR}/references/errors.md for comprehensive error handling.
See ${CLAUDE_SKILL_DIR}/references/examples.md for detailed examples.
${CLAUDE_SKILL_DIR}/scripts/validate_config.py${CLAUDE_SKILL_DIR}/scripts/render_template.py${CLAUDE_SKILL_DIR}/scripts/quality_score.py${CLAUDE_SKILL_DIR}/assets/default-changelog.md${CLAUDE_SKILL_DIR}/assets/weekly-template.md${CLAUDE_SKILL_DIR}/assets/release-template.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 3,396 | 8,215 | +142% | 1 | 1 | 0% | 312 | 1,097 | +252% | 0 | 0 | — |
case-02 | fail→fail | 6,154 | 7,105 | +15% | 1 | 1 | 0% | 537 | 1,072 | +100% | 0 | 0 | — |
case-03 | fail→fail | 3,762 | 5,945 | +58% | 1 | 1 | 0% | 213 | 1,021 | +379% | 0 | 0 | — |
case-04 | pass→pass | 10,100 | 7,829 | -22% | 1 | 1 | 0% | 1,887 | 2,147 | +14% | 0 | 0 | — |
case-05 | pass→pass | 3,403 | 3,156 | -7% | 1 | 1 | 0% | 540 | 1,160 | +115% | 0 | 0 | — |
case-06 | pass→pass | 7,784 | 6,214 | -20% | 1 | 1 | 0% | 1,362 | 1,643 | +21% | 0 | 0 | — |
case-07 | fail→pass | 6,410 | 2,380 | -63% | 1 | 1 | 0% | 1,068 | 979 | -8% | 0 | 0 | — |
case-08 | fail→pass | 7,718 | 2,259 | -71% | 1 | 1 | 0% | 1,347 | 868 | -36% | 0 | 0 | — |
case-09 | fail→pass | 7,678 | 3,437 | -55% | 1 | 1 | 0% | 1,172 | 786 | -33% | 0 | 0 | — |
case-10 | pass→pass | 3,789 | 2,403 | -37% | 1 | 1 | 0% | 619 | 929 | +50% | 0 | 0 | — |
case-11 | pass→pass | 6,221 | 1,904 | -69% | 1 | 1 | 0% | 1,009 | 821 | -19% | 0 | 0 | — |
case-12 | fail→pass | 8,770 | 1,726 | -80% | 1 | 1 | 0% | 1,359 | 837 | -38% | 0 | 0 | — |
case-13 | pass→pass | 7,662 | 2,899 | -62% | 1 | 1 | 0% | 1,355 | 1,072 | -21% | 0 | 0 | — |
case-14 | fail→pass | 6,574 | 1,754 | -73% | 1 | 1 | 0% | 1,070 | 878 | -18% | 0 | 0 | — |
case-15 | pass→pass | 6,543 | 2,087 | -68% | 1 | 1 | 0% | 1,110 | 910 | -18% | 0 | 0 | — |
case-16 | fail→pass | 9,413 | 3,751 | -60% | 1 | 1 | 0% | 1,745 | 1,270 | -27% | 0 | 0 | — |
case-17 | fail→pass | 6,167 | 7,677 | +24% | 1 | 1 | 0% | 963 | 1,799 | +87% | 0 | 0 | — |
case-18 | fail→pass | 4,754 | 2,367 | -50% | 1 | 1 | 0% | 673 | 855 | +27% | 0 | 0 | — |
case-19 | pass→pass | 5,033 | 1,439 | -71% | 1 | 1 | 0% | 724 | 747 | +3% | 0 | 0 | — |
case-20 | fail→pass | 6,736 | 1,826 | -73% | 1 | 1 | 0% | 986 | 869 | -12% | 0 | 0 | — |
case-21 | fail→pass | 5,995 | 1,646 | -73% | 1 | 1 | 0% | 845 | 787 | -7% | 0 | 0 | — |
case-22 | fail→pass | 8,178 | 1,532 | -81% | 1 | 1 | 0% | 1,203 | 842 | -30% | 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 19 counted toward the lift figure. The other 3 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 +50 percentage points is the difference between those two pass rates over the 19 comparable cases.
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.