Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Audit milestone completion against original intent before archiving
.claude/skills/coco-research-gsd-audit-milestone/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -84% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -84% | 0% |
<objective> Verify milestone achieved its definition of done. Check requirements coverage, cross-phase integration, and end-to-end flows.
This command IS the orchestrator. Reads existing VERIFICATION.md files (phases already verified during execute-phase), aggregates tech debt and deferred gaps, then spawns integration checker for cross-phase wiring. </objective>
<execution_context> @$HOME/.claude/get-shit-done/workflows/audit-milestone.md </execution_context>
<context> Version: $ARGUMENTS (optional — defaults to current milestone)
Core planning files are resolved in-workflow (init milestone-op) and loaded only as needed.
Completed Work: Glob: .planning/phases//-SUMMARY.md Glob: .planning/phases//-VERIFICATION.md </context>
<process> Execute the audit-milestone workflow from @$HOME/.claude/get-shit-done/workflows/audit-milestone.md end-to-end. Preserve all workflow gates (scope determination, verification reading, integration check, requirements coverage, routing). </process>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | fail→fail | 14,450 | 14,364 | -1% | 1 | 1 | 0% | 1,489 | 1,653 | +11% | 0 | 0 | — |
case-01 | pass→fail | 35,906 | 11,633 | -68% | 1 | 1 | 0% | 5,077 | 526 | -90% | 0 | 0 | — |
case-02 | fail→fail | 17,138 | 15,537 | -9% | 1 | 1 | 0% | 2,077 | 398 | -81% | 0 | 0 | — |
case-03 | fail→fail | 33,427 | 4,736 | -86% | 1 | 1 | 0% | 4,841 | 470 | -90% | 0 | 0 | — |
case-04 | pass→fail | 13,226 | 15,665 | +18% | 1 | 1 | 0% | 1,349 | 589 | -56% | 0 | 0 | — |
case-05 | pass→fail | 18,015 | 8,072 | -55% | 1 | 1 | 0% | 2,905 | 632 | -78% | 0 | 0 | — |
case-06 | pass→fail | 25,950 | 10,419 | -60% | 1 | 1 | 0% | 3,083 | 1,017 | -67% | 0 | 0 | — |
case-07 | fail→fail | 11,348 | 12,032 | +6% | 1 | 1 | 0% | 1,685 | 632 | -62% | 0 | 0 | — |
case-08 | fail→fail | 16,487 | 20,296 | +23% | 1 | 1 | 0% | 1,728 | 764 | -56% | 0 | 0 | — |
case-09 | fail→pass | 40,796 | 1,632 | -96% | 1 | 1 | 0% | 2,900 | 470 | -84% | 0 | 0 | — |
case-10 | fail→pass | 12,470 | 6,905 | -45% | 1 | 1 | 0% | 1,931 | 1,189 | -38% | 0 | 0 | — |
case-11 | fail→fail | 20,116 | 11,929 | -41% | 1 | 1 | 0% | 2,207 | 1,424 | -35% | 0 | 0 | — |
case-12 | pass→pass | 18,542 | 14,760 | -20% | 1 | 1 | 0% | 1,874 | 1,071 | -43% | 0 | 0 | — |
case-13 | fail→pass | 16,053 | 6,924 | -57% | 1 | 1 | 0% | 1,510 | 526 | -65% | 0 | 0 | — |
case-14 | fail→fail | 17,384 | 7,965 | -54% | 1 | 1 | 0% | 1,903 | 713 | -63% | 0 | 0 | — |
case-15 | fail→fail | 21,014 | 13,595 | -35% | 1 | 1 | 0% | 2,312 | 847 | -63% | 0 | 0 | — |
case-16 | fail→pass | 14,198 | 7,291 | -49% | 1 | 1 | 0% | 1,569 | 551 | -65% | 0 | 0 | — |
case-17 | pass→pass | 15,489 | 13,040 | -16% | 1 | 1 | 0% | 1,669 | 1,566 | -6% | 0 | 0 | — |
case-18 | fail→pass | 18,239 | 2,664 | -85% | 1 | 1 | 0% | 3,468 | 545 | -84% | 0 | 0 | — |
case-20 | fail→fail | 13,878 | 10,202 | -26% | 1 | 1 | 0% | 1,418 | 762 | -46% | 0 | 0 | — |
case-21 | fail→pass | 14,617 | 11,887 | -19% | 1 | 1 | 0% | 1,444 | 1,277 | -12% | 0 | 0 | — |
case-22 | fail→pass | 21,070 | 11,064 | -47% | 1 | 1 | 0% | 2,369 | 1,177 | -50% | 0 | 0 | — |
case-23 | pass→fail | 12,666 | 6,832 | -46% | 1 | 1 | 0% | 1,657 | 742 | -55% | 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. 23 cases were attempted, and 12 counted toward the lift figure. The other 11 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 +9 percentage points is the difference between those two pass rates over the 12 comparable cases. 7 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.