Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when maintaining or releasing a public open-source repository, including issue triage, pull request review, dependency or security remediation, release preparation, and source-linked adoption reporting.
.claude/skills/hashgraph-online-oss-maintainer-workflow/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 106% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -6% | 0% |
Run one small, verifiable maintenance cycle for a public repository. The goal is not activity volume. The goal is a real maintainer outcome backed by public issues, pull requests, CI, security checks, release notes, and honest adoption evidence.
docs/oss-maintenance/<cycle>/01-intake-and-signals.mddocs/oss-maintenance/<cycle>/02-maintenance-plan.mddocs/oss-maintenance/<cycle>/03-change-and-review-evidence.mddocs/oss-maintenance/<cycle>/04-release-readiness.mddocs/oss-maintenance/<cycle>/05-adoption-follow-up.mdplan_oss_maintenance_cyclereview_oss_release_readinessCI, dependency audit, and security policy.
plan_oss_maintenance_cycle with source-backed counts. Do not estimatestars, downloads, users, or vulnerabilities.
review changes outrank speculative features.
issue solely to make the repository look active.
build, validation, and security audit.
rollback path in the maintenance evidence.
CHANGELOG.md.review_oss_release_readiness. Do not tag or publish while it reportsblocked.
a primary or core maintainer.
maintainer-requested fix is available.
sustained review and release responsibility.
review_oss_release_readiness.nextAllowedActionRead references/maintenance-evidence-contract.md before preparing the release gate or reporting ecosystem and adoption signals.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 58,836 | 46,691 | -21% | 1 | 1 | 0% | 8,330 | 9,137 | +10% | 0 | 0 | — |
case-02 | fail→fail | 17,004 | 14,058 | -17% | 1 | 1 | 0% | 259 | 1,162 | +349% | 0 | 0 | — |
case-03 | fail→fail | 53,231 | 51,806 | -3% | 1 | 1 | 0% | 8,331 | 9,138 | +10% | 0 | 0 | — |
case-04 | pass→pass | 15,470 | 21,334 | +38% | 1 | 1 | 0% | 2,461 | 4,716 | +92% | 0 | 0 | — |
case-05 | pass→pass | 10,432 | 16,884 | +62% | 1 | 1 | 0% | 1,612 | 3,386 | +110% | 0 | 0 | — |
case-06 | pass→pass | 21,931 | 20,426 | -7% | 1 | 1 | 0% | 2,764 | 4,180 | +51% | 0 | 0 | — |
case-07 | pass→fail | 21,929 | 42,825 | +95% | 1 | 1 | 0% | 2,009 | 1,300 | -35% | 0 | 0 | — |
case-08 | fail→pass | 11,970 | 14,428 | +21% | 1 | 1 | 0% | 1,700 | 2,468 | +45% | 0 | 0 | — |
case-09 | fail→pass | 12,901 | 12,911 | +0% | 1 | 1 | 0% | 1,401 | 2,892 | +106% | 0 | 0 | — |
case-10 | pass→pass | 10,095 | 24,735 | +145% | 1 | 1 | 0% | 996 | 3,335 | +235% | 0 | 0 | — |
case-11 | pass→pass | 14,669 | 34,574 | +136% | 1 | 1 | 0% | 2,312 | 4,266 | +85% | 0 | 0 | — |
case-12 | pass→pass | 21,627 | 8,725 | -60% | 1 | 1 | 0% | 786 | 1,521 | +94% | 0 | 0 | — |
case-13 | fail→pass | 19,481 | 27,170 | +39% | 1 | 1 | 0% | 1,997 | 2,482 | +24% | 0 | 0 | — |
case-14 | pass→pass | 12,720 | 7,079 | -44% | 1 | 1 | 0% | 1,550 | 1,839 | +19% | 0 | 0 | — |
case-15 | pass→pass | 14,024 | 10,285 | -27% | 1 | 1 | 0% | 1,446 | 1,620 | +12% | 0 | 0 | — |
case-16 | fail→pass | 22,295 | 18,604 | -17% | 1 | 1 | 0% | 3,605 | 3,308 | -8% | 0 | 0 | — |
case-17 | fail→pass | 12,971 | 8,459 | -35% | 1 | 1 | 0% | 1,485 | 1,391 | -6% | 0 | 0 | — |
case-18 | pass→pass | 12,380 | 9,329 | -25% | 1 | 1 | 0% | 1,034 | 1,708 | +65% | 0 | 0 | — |
case-19 | pass→pass | 11,156 | 22,070 | +98% | 1 | 1 | 0% | 1,552 | 3,364 | +117% | 0 | 0 | — |
case-20 | pass→pass | 11,598 | 12,993 | +12% | 1 | 1 | 0% | 1,879 | 1,967 | +5% | 0 | 0 | — |
case-21 | fail→pass | 14,725 | 6,539 | -56% | 1 | 1 | 0% | 1,473 | 1,293 | -12% | 0 | 0 | — |
case-22 | pass→pass | 15,569 | 3,523 | -77% | 1 | 1 | 0% | 1,555 | 1,405 | -10% | 0 | 0 | — |
case-23 | fail→pass | 10,441 | 7,842 | -25% | 1 | 1 | 0% | 1,653 | 971 | -41% | 0 | 0 | — |
case-24 | pass→pass | 15,632 | 12,857 | -18% | 1 | 1 | 0% | 1,519 | 1,864 | +23% | 0 | 0 | — |
case-25 | pass→pass | 19,391 | 21,764 | +12% | 1 | 1 | 0% | 2,099 | 1,625 | -23% | 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. 25 cases were attempted, and 23 counted toward the lift figure. The other 2 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 +24 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is 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.