Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate tests for a completed phase based on UAT criteria and implementation
.claude/skills/davepoon-gsd-add-tests/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 30% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -67% | 0% |
| case-20 | ✓→✗ | ▼ Worse | -77% | 0% |
<objective> Generate unit and E2E tests for a completed phase, using its SUMMARY.md, CONTEXT.md, and VERIFICATION.md as specifications.
Analyzes implementation files, classifies them into TDD (unit), E2E (browser), or Skip categories, presents a test plan for user approval, then generates tests following RED-GREEN conventions.
Output: Test files committed with message test(phase-{N}): add unit and E2E tests from add-tests command </objective>
<execution_context> @${CLAUDE_PLUGIN_ROOT}/workflows/add-tests.md </execution_context>
<context> Phase: $ARGUMENTS
@.planning/STATE.md @.planning/ROADMAP.md </context>
<process> Execute the add-tests workflow from @${CLAUDE_PLUGIN_ROOT}/workflows/add-tests.md end-to-end. Preserve all workflow gates (classification approval, test plan approval, RED-GREEN verification, gap reporting). </process>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 2,527 | 4,872 | +93% | 1 | 1 | 0% | 356 | 552 | +55% | 0 | 0 | — |
case-02 | fail→fail | 7,105 | 4,309 | -39% | 1 | 1 | 0% | 160 | 458 | +186% | 0 | 0 | — |
case-03 | fail→fail | 7,268 | 4,105 | -44% | 1 | 1 | 0% | 1,310 | 453 | -65% | 0 | 0 | — |
case-04 | fail→fail | 8,832 | 4,407 | -50% | 1 | 1 | 0% | 1,626 | 449 | -72% | 0 | 0 | — |
case-05 | fail→fail | 3,775 | 4,508 | +19% | 1 | 1 | 0% | 625 | 386 | -38% | 0 | 0 | — |
case-06 | fail→fail | 21,068 | 4,384 | -79% | 1 | 1 | 0% | 4,807 | 407 | -92% | 0 | 0 | — |
case-07 | fail→fail | 5,168 | 4,860 | -6% | 1 | 1 | 0% | 849 | 430 | -49% | 0 | 0 | — |
case-08 | fail→pass | 9,738 | 2,972 | -69% | 1 | 1 | 0% | 1,678 | 573 | -66% | 0 | 0 | — |
case-09 | fail→pass | 7,709 | 2,880 | -63% | 1 | 1 | 0% | 1,416 | 711 | -50% | 0 | 0 | — |
case-10 | pass→pass | 6,376 | 2,769 | -57% | 1 | 1 | 0% | 1,232 | 719 | -42% | 0 | 0 | — |
case-11 | fail→pass | 5,840 | 11,648 | +99% | 1 | 1 | 0% | 958 | 1,243 | +30% | 0 | 0 | — |
case-12 | fail→fail | 8,301 | 4,530 | -45% | 1 | 1 | 0% | 1,390 | 392 | -72% | 0 | 0 | — |
case-13 | fail→fail | 10,840 | 6,014 | -45% | 1 | 1 | 0% | 2,048 | 456 | -78% | 0 | 0 | — |
case-14 | pass→pass | 9,540 | 7,028 | -26% | 1 | 1 | 0% | 1,551 | 1,189 | -23% | 0 | 0 | — |
case-15 | pass→pass | 5,693 | 4,850 | -15% | 1 | 1 | 0% | 853 | 1,035 | +21% | 0 | 0 | — |
case-16 | pass→pass | 8,536 | 3,873 | -55% | 1 | 1 | 0% | 1,459 | 842 | -42% | 0 | 0 | — |
case-17 | pass→pass | 10,139 | 6,647 | -34% | 1 | 1 | 0% | 1,758 | 1,372 | -22% | 0 | 0 | — |
case-18 | fail→pass | 7,696 | 1,438 | -81% | 1 | 1 | 0% | 1,356 | 451 | -67% | 0 | 0 | — |
case-19 | pass→pass | 10,793 | 4,418 | -59% | 1 | 1 | 0% | 1,927 | 917 | -52% | 0 | 0 | — |
case-20 | pass→fail | 10,059 | 7,051 | -30% | 1 | 1 | 0% | 2,314 | 541 | -77% | 0 | 0 | — |
case-21 | pass→fail | 10,750 | 4,536 | -58% | 1 | 1 | 0% | 2,476 | 458 | -82% | 0 | 0 | — |
case-22 | pass→fail | 14,344 | 3,709 | -74% | 1 | 1 | 0% | 2,672 | 472 | -82% | 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 10 counted toward the lift figure. The other 12 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 10 comparable cases. 4 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.