Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Detect missing test coverage and generate test suggestions. Use when the user asks about coverage gaps, untested code, or what tests to write next; also after adding a feature to find what still needs tests.
.claude/skills/ruvnet-test-gaps/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-11 | ✗→✓ | ▲ Improved | -73% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -75% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -53% | 0% |
Find test coverage gaps via CLI:
bashnpx @claude-flow/cli@latest hooks coverage-gaps --format table --limit 20 npx @claude-flow/cli@latest hooks coverage-route --task "add auth tests" npx @claude-flow/cli@latest hooks coverage-suggest --path src/
Or dispatch the testgaps worker via MCP: mcp__plugin_ruflo-core_ruflo__hooks_worker-dispatch({ trigger: "testgaps" })
For continuous detection, use /loop with the loop-worker skill targeting the testgaps worker.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-11 | fail→pass | 16,233 | 3,325 | -80% | 1 | 1 | 0% | 2,850 | 763 | -73% | 0 | 0 | — |
case-01 | fail→fail | 4,772 | 9,861 | +107% | 1 | 1 | 0% | 808 | 1,133 | +40% | 0 | 0 | — |
case-02 | fail→pass | 9,670 | 7,220 | -25% | 1 | 1 | 0% | 1,746 | 1,568 | -10% | 0 | 0 | — |
case-03 | fail→pass | 13,339 | 5,394 | -60% | 1 | 1 | 0% | 2,552 | 1,276 | -50% | 0 | 0 | — |
case-04 | fail→pass | 12,357 | 2,277 | -82% | 1 | 1 | 0% | 2,468 | 625 | -75% | 0 | 0 | — |
case-05 | fail→pass | 14,896 | 4,861 | -67% | 1 | 1 | 0% | 2,312 | 1,090 | -53% | 0 | 0 | — |
case-06 | fail→pass | 3,521 | 2,223 | -37% | 1 | 1 | 0% | 558 | 523 | -6% | 0 | 0 | — |
case-07 | fail→pass | 13,121 | 3,035 | -77% | 1 | 1 | 0% | 2,692 | 733 | -73% | 0 | 0 | — |
case-08 | fail→pass | 13,772 | 4,933 | -64% | 1 | 1 | 0% | 2,304 | 980 | -57% | 0 | 0 | — |
case-09 | fail→pass | 7,756 | 1,930 | -75% | 1 | 1 | 0% | 1,426 | 464 | -67% | 0 | 0 | — |
case-10 | fail→pass | 7,342 | 1,784 | -76% | 1 | 1 | 0% | 1,304 | 421 | -68% | 0 | 0 | — |
case-12 | fail→pass | 11,269 | 3,019 | -73% | 1 | 1 | 0% | 2,329 | 695 | -70% | 0 | 0 | — |
case-13 | fail→pass | 14,295 | 3,781 | -74% | 1 | 1 | 0% | 2,559 | 862 | -66% | 0 | 0 | — |
case-14 | fail→pass | 8,264 | 1,717 | -79% | 1 | 1 | 0% | 1,347 | 419 | -69% | 0 | 0 | — |
case-15 | fail→pass | 14,521 | 3,055 | -79% | 1 | 1 | 0% | 2,624 | 546 | -79% | 0 | 0 | — |
case-16 | fail→pass | 5,279 | 2,203 | -58% | 1 | 1 | 0% | 988 | 451 | -54% | 0 | 0 | — |
case-17 | fail→pass | 7,043 | 7,395 | +5% | 1 | 1 | 0% | 1,325 | 468 | -65% | 0 | 0 | — |
case-18 | fail→pass | 10,050 | 2,032 | -80% | 1 | 1 | 0% | 1,701 | 395 | -77% | 0 | 0 | — |
case-19 | fail→pass | 16,716 | 1,579 | -91% | 1 | 1 | 0% | 2,145 | 457 | -79% | 0 | 0 | — |
case-20 | pass→pass | 11,068 | 11,235 | +2% | 1 | 1 | 0% | 2,771 | 2,900 | +5% | 0 | 0 | — |
case-21 | pass→pass | 6,356 | 5,945 | -6% | 1 | 1 | 0% | 1,395 | 1,381 | -1% | 0 | 0 | — |
case-22 | pass→pass | 7,889 | 6,629 | -16% | 1 | 1 | 0% | 1,594 | 1,602 | +1% | 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 +82 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.