Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Auto-detect testing framework and run relevant tests. Identifies Jest, Vitest, Playwright, Cypress, pytest, Go test, and others. Can run all tests, specific file tests, or generate basic tests for new code. Usage - /test-detect, /test-detect src/auth/login.ts, /test-detect generate src/utils.ts
.claude/skills/davila7-test-detect/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | 55% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 125% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-24 | ✓→✗ | ▼ Worse | -21% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 31% | 0% |
Automatically detect the testing framework in the current project and run the right tests.
Check for these files in order (first match wins):
| Check | Framework | Run Command | |-------|-----------|-------------| | vitest.config.* exists OR vitest in devDeps | Vitest | npx vitest run | | jest.config.* exists OR jest in devDeps | Jest | npx jest | | playwright.config.* exists | Playwright | npx playwright test | | cypress.config.* exists | Cypress | npx cypress run | | pytest.ini or conftest.py or pyproject.toml with [tool.pytest] | pytest | python -m pytest | | go.mod exists | Go test | go test ./... | | Cargo.toml exists | Rust/cargo | cargo test | | mix.exs exists | ExUnit | mix test | | Gemfile with rspec | RSpec | bundle exec rspec | | package.json has scripts.test | npm test | npm test |
Report the detected framework before proceeding.
Check $ARGUMENTS for the mode:
src/auth/login.ts): Run tests for that file (Step 4)generate src/utils.ts): Generate tests (Step 5)Run the detected test command. After completion:
/test-detect <failing-file> to investigate a specific failure"Given a source file path, find its test file:
Search strategy (try in order):
__tests__/<filename>.test.<ext> (Jest convention)<filename>.test.<ext> (co-located)<filename>.spec.<ext> (alternative convention)test/<filename>_test.<ext> (Go/Python convention)tests/test_<filename>.<ext> (pytest convention)<filename>_test.go (Go convention)Use Glob to find matches. If found, run only that test file:
| Framework | Command | |-----------|---------| | Vitest | npx vitest run <test-file> | | Jest | npx jest <test-file> | | Playwright | npx playwright test <test-file> | | pytest | python -m pytest <test-file> | | Go | go test -run <TestName> ./<package>/ | | Cargo | cargo test <test_name> |
If no test file found, ask: "No tests found for this file. Want me to generate them? Run /test-detect generate <file>"
Read the source file and analyze:
describe block per function/classit/test blocks covering: happy path, edge cases, error casesSave to the conventional location for the detected framework:
__tests__/<filename>.test.<ext> or <filename>.test.<ext> (match existing convention)tests/test_<filename>.py<filename>_test.go (same directory)spec/<filename>_spec.rbShow the generated file path and ask if the user wants to run the new tests.
package.json has both test and test:unit/test:e2e scripts, prefer the specific one when context is clear| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 8,341 | 8,686 | +4% | 1 | 1 | 0% | 160 | 1,379 | +762% | 0 | 0 | — |
case-02 | fail→fail | 12,278 | 2,593 | -79% | 1 | 1 | 0% | 2,862 | 1,418 | -50% | 0 | 0 | — |
case-03 | pass→pass | 8,730 | 5,427 | -38% | 1 | 1 | 0% | 1,694 | 2,224 | +31% | 0 | 0 | — |
case-04 | pass→pass | 4,588 | 2,530 | -45% | 1 | 1 | 0% | 908 | 1,458 | +61% | 0 | 0 | — |
case-05 | pass→pass | 2,539 | 2,242 | -12% | 1 | 1 | 0% | 418 | 1,473 | +252% | 0 | 0 | — |
case-06 | pass→pass | 2,854 | 1,733 | -39% | 1 | 1 | 0% | 489 | 1,312 | +168% | 0 | 0 | — |
case-07 | pass→pass | 2,871 | 2,360 | -18% | 1 | 1 | 0% | 484 | 1,436 | +197% | 0 | 0 | — |
case-08 | pass→pass | 1,249 | 1,856 | +49% | 1 | 1 | 0% | 172 | 1,313 | +663% | 0 | 0 | — |
case-09 | pass→pass | 3,013 | 1,404 | -53% | 1 | 1 | 0% | 499 | 1,278 | +156% | 0 | 0 | — |
case-10 | fail→pass | 5,125 | 2,560 | -50% | 1 | 1 | 0% | 1,043 | 1,614 | +55% | 0 | 0 | — |
case-11 | pass→pass | 4,066 | 2,612 | -36% | 1 | 1 | 0% | 705 | 1,331 | +89% | 0 | 0 | — |
case-12 | pass→pass | 2,175 | 1,762 | -19% | 1 | 1 | 0% | 354 | 1,362 | +285% | 0 | 0 | — |
case-13 | fail→pass | 3,704 | 2,724 | -26% | 1 | 1 | 0% | 626 | 1,409 | +125% | 0 | 0 | — |
case-14 | fail→fail | 4,752 | 1,832 | -61% | 1 | 1 | 0% | 804 | 1,371 | +71% | 0 | 0 | — |
case-15 | fail→pass | 7,663 | 1,395 | -82% | 1 | 1 | 0% | 1,354 | 1,253 | -7% | 0 | 0 | — |
case-16 | fail→fail | 4,275 | 3,206 | -25% | 1 | 1 | 0% | 860 | 1,614 | +88% | 0 | 0 | — |
case-17 | pass→pass | 5,175 | 2,518 | -51% | 1 | 1 | 0% | 1,002 | 1,460 | +46% | 0 | 0 | — |
case-18 | fail→fail | 2,647 | 2,238 | -15% | 1 | 1 | 0% | 421 | 1,497 | +256% | 0 | 0 | — |
case-19 | pass→pass | 8,327 | 3,121 | -63% | 1 | 1 | 0% | 1,525 | 1,656 | +9% | 0 | 0 | — |
case-20 | pass→pass | 4,072 | 2,640 | -35% | 1 | 1 | 0% | 813 | 1,657 | +104% | 0 | 0 | — |
case-21 | pass→pass | 14,068 | 7,047 | -50% | 1 | 1 | 0% | 2,847 | 2,527 | -11% | 0 | 0 | — |
case-22 | pass→pass | 8,547 | 8,726 | +2% | 1 | 1 | 0% | 1,890 | 2,970 | +57% | 0 | 0 | — |
case-23 | pass→pass | 9,890 | 6,939 | -30% | 1 | 1 | 0% | 1,993 | 2,601 | +31% | 0 | 0 | — |
case-24 | pass→fail | 8,649 | 4,543 | -47% | 1 | 1 | 0% | 1,722 | 1,366 | -21% | 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. 24 cases were attempted, and 22 counted toward the lift figure. The other 2 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 +8 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.