Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyze phase dependencies and suggest Depends on entries for ROADMAP.md
.claude/skills/davepoon-gsd-analyze-dependencies/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-21 | ✓→✗ | ▼ Worse | 74% | 0% |
<objective> Analyze the phase dependency graph for the current milestone. For each phase pair, determine if there is a dependency relationship based on:
Then suggest Depends on updates to ROADMAP.md. </objective>
<execution_context> @${CLAUDE_PLUGIN_ROOT}/workflows/analyze-dependencies.md </execution_context>
<context> No arguments required. Requires an active milestone with ROADMAP.md.
Run this command BEFORE /gsd:manager to fill in missing Depends on fields and prevent merge conflicts from unordered parallel execution. </context>
<process> Execute the analyze-dependencies workflow from @${CLAUDE_PLUGIN_ROOT}/workflows/analyze-dependencies.md end-to-end. Present dependency suggestions clearly and apply confirmed updates to ROADMAP.md. </process>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-21 | pass→fail | 5,026 | 4,797 | -5% | 1 | 1 | 0% | 335 | 583 | +74% | 0 | 0 | — |
case-01 | fail→fail | 7,112 | 3,646 | -49% | 1 | 1 | 0% | 1,182 | 464 | -61% | 0 | 0 | — |
case-02 | fail→fail | 4,620 | 3,389 | -27% | 1 | 1 | 0% | 713 | 398 | -44% | 0 | 0 | — |
case-03 | fail→fail | 3,949 | 3,655 | -7% | 1 | 1 | 0% | 273 | 410 | +50% | 0 | 0 | — |
case-04 | pass→pass | 7,280 | 3,490 | -52% | 1 | 1 | 0% | 1,371 | 849 | -38% | 0 | 0 | — |
case-05 | pass→pass | 7,742 | 2,525 | -67% | 1 | 1 | 0% | 1,580 | 675 | -57% | 0 | 0 | — |
case-06 | pass→pass | 1,946 | 2,242 | +15% | 1 | 1 | 0% | 325 | 586 | +80% | 0 | 0 | — |
case-07 | pass→pass | 5,155 | 1,437 | -72% | 1 | 1 | 0% | 874 | 451 | -48% | 0 | 0 | — |
case-08 | fail→pass | 7,818 | 2,927 | -63% | 1 | 1 | 0% | 1,451 | 772 | -47% | 0 | 0 | — |
case-09 | pass→pass | 3,956 | 1,976 | -50% | 1 | 1 | 0% | 695 | 620 | -11% | 0 | 0 | — |
case-10 | pass→pass | 5,958 | 3,336 | -44% | 1 | 1 | 0% | 1,197 | 798 | -33% | 0 | 0 | — |
case-11 | pass→fail | 10,859 | 3,168 | -71% | 1 | 1 | 0% | 1,957 | 490 | -75% | 0 | 0 | — |
case-12 | pass→pass | 6,042 | 2,459 | -59% | 1 | 1 | 0% | 1,088 | 608 | -44% | 0 | 0 | — |
case-13 | fail→fail | 10,103 | 4,759 | -53% | 1 | 1 | 0% | 1,668 | 980 | -41% | 0 | 0 | — |
case-14 | pass→pass | 3,320 | 2,670 | -20% | 1 | 1 | 0% | 608 | 650 | +7% | 0 | 0 | — |
case-15 | pass→pass | 5,188 | 2,618 | -50% | 1 | 1 | 0% | 887 | 659 | -26% | 0 | 0 | — |
case-16 | pass→pass | 3,241 | 3,340 | +3% | 1 | 1 | 0% | 582 | 797 | +37% | 0 | 0 | — |
case-17 | fail→pass | 7,695 | 2,738 | -64% | 1 | 1 | 0% | 1,322 | 702 | -47% | 0 | 0 | — |
case-18 | fail→pass | 7,242 | 2,724 | -62% | 1 | 1 | 0% | 1,469 | 704 | -52% | 0 | 0 | — |
case-19 | fail→pass | 7,220 | 4,244 | -41% | 1 | 1 | 0% | 1,163 | 868 | -25% | 0 | 0 | — |
case-20 | pass→fail | 12,411 | 4,918 | -60% | 1 | 1 | 0% | 2,328 | 456 | -80% | 0 | 0 | — |
case-22 | pass→pass | 2,899 | 2,971 | +2% | 1 | 1 | 0% | 499 | 745 | +49% | 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 16 counted toward the lift figure. The other 6 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 +5 percentage points is the difference between those two pass rates over the 16 comparable cases. 3 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.