Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Review branch changes against Archcore docs, or report project health. Use for 'review my branch', 'review the changes before merge', 'show status', 'documentation gaps', 'check if docs match code', 'close out the feature', 'ship the feature and close it out', or after a staleness warning. --drift for staleness detection, --deep for a full documentation audit. Not for creating docs — use /archcore:document; not for planning — use /archcore:plan.
.claude/skills/hashgraph-online-review/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 49% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -11% | 0% |
If topology=multi-module (triage announcement or coordinator spec/plan declaration), read ../references/multi-module.md. Review each module diff in its owning repository first, then review the workspace as one change set. Block missing repositories, copied or divergent shared contracts, undeclared cross-module edits, overlapping writers, incomplete local acceptance, or integration results that do not identify the tested workspace combination.
Check in order:
convention, idiomatic for the stack? (see ../references/quality.md)contract covers full public surface? Env vars/config/error modes documented? README commands runnable? Living docs honor North Star (atemporal ground truth, modular specs; see ../references/quality.md)?For spec/plan reviews, also block unresolved implementation notes, plan assumptions absent from spec, vague acceptance, or premature [parallel].
Report:
BLOCK: <must fix>
FIX: <should fix>
NIT: <optional>BLOCKs resolved before merge. FIX resolved or explicitly deferred with reason.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | fail→fail | 49,889 | 52,121 | +4% | 1 | 1 | 0% | 8,106 | 7,454 | -8% | 0 | 0 | — |
case-01 | fail→fail | 13,067 | 7,741 | -41% | 1 | 1 | 0% | 1,110 | 748 | -33% | 0 | 0 | — |
case-02 | fail→fail | 9,487 | 14,270 | +50% | 1 | 1 | 0% | 1,151 | 613 | -47% | 0 | 0 | — |
case-03 | fail→fail | 9,007 | 12,366 | +37% | 1 | 1 | 0% | 1,367 | 1,391 | +2% | 0 | 0 | — |
case-04 | fail→fail | 16,186 | 23,290 | +44% | 1 | 1 | 0% | 2,388 | 2,206 | -8% | 0 | 0 | — |
case-05 | fail→fail | 12,841 | 7,012 | -45% | 1 | 1 | 0% | 1,226 | 1,505 | +23% | 0 | 0 | — |
case-07 | fail→pass | 15,668 | 9,071 | -42% | 1 | 1 | 0% | 2,344 | 1,838 | -22% | 0 | 0 | — |
case-08 | fail→pass | 17,613 | 7,074 | -60% | 1 | 1 | 0% | 2,710 | 1,324 | -51% | 0 | 0 | — |
case-09 | fail→pass | 12,493 | 4,066 | -67% | 1 | 1 | 0% | 1,693 | 1,058 | -38% | 0 | 0 | — |
case-10 | fail→pass | 14,020 | 11,770 | -16% | 1 | 1 | 0% | 929 | 1,384 | +49% | 0 | 0 | — |
case-11 | fail→pass | 8,684 | 3,734 | -57% | 1 | 1 | 0% | 1,147 | 1,019 | -11% | 0 | 0 | — |
case-16 | fail→pass | 15,951 | 13,009 | -18% | 1 | 1 | 0% | 1,781 | 1,492 | -16% | 0 | 0 | — |
case-12 | fail→pass | 14,606 | 4,711 | -68% | 1 | 1 | 0% | 1,927 | 1,012 | -47% | 0 | 0 | — |
case-13 | pass→pass | 14,126 | 7,685 | -46% | 1 | 1 | 0% | 2,228 | 1,468 | -34% | 0 | 0 | — |
case-14 | fail→pass | 18,142 | 9,479 | -48% | 1 | 1 | 0% | 1,921 | 1,074 | -44% | 0 | 0 | — |
case-15 | fail→pass | 15,629 | 5,265 | -66% | 1 | 1 | 0% | 2,226 | 1,146 | -49% | 0 | 0 | — |
case-17 | pass→pass | 11,583 | 5,771 | -50% | 1 | 1 | 0% | 1,920 | 1,354 | -29% | 0 | 0 | — |
case-18 | pass→pass | 11,646 | 9,200 | -21% | 1 | 1 | 0% | 1,639 | 1,057 | -36% | 0 | 0 | — |
case-19 | pass→pass | 18,470 | 22,060 | +19% | 1 | 1 | 0% | 3,020 | 3,049 | +1% | 0 | 0 | — |
case-20 | pass→pass | 16,282 | 7,241 | -56% | 1 | 1 | 0% | 1,839 | 1,577 | -14% | 0 | 0 | — |
case-21 | pass→pass | 12,932 | 9,547 | -26% | 1 | 1 | 0% | 2,221 | 1,219 | -45% | 0 | 0 | — |
case-22 | pass→pass | 16,235 | 5,195 | -68% | 1 | 1 | 0% | 2,147 | 1,252 | -42% | 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 21 counted toward the lift figure. The other 1 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 +41 percentage points is the difference between those two pass rates over the 21 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.