- remove legacy .ai/ knowledge shots and reports, semantic skills invariant assessment, and obsolete .kilo/workflows/ - update agent model selection (omniroute/terra) for qa-tester, security-auditor, svelte-coder - add swarm-master agent and speckit openapi/prototype/resume/validate plus test.* commands; update speckit plan/ux/implement docs - refresh skill SKILL.md files (semantics core/testing/svelte/belief, molecular-cot-logging, semantic-frontend) - add semantic curation report
3.1 KiB
description, handoffs, tools
| description | handoffs | tools | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Run full test suite: backend unit tests, frontend vitest, coverage reports. Use as final verification gate. |
|
bash, grep, read |
User Input
$ARGUMENTS
Goal
Run the COMPLETE test suite across both backend and frontend, producing a unified pass/fail + coverage report. This is the final verification gate before code review or merge.
Required Skills
MANDATORY USE skill({name="semantics-testing"}) — test conventions, anti-tautology rules, tier markers.
MANDATORY USE skill({name="molecular-cot-logging"}) — structured logging during execution.
Execution Steps
1. Pre-flight checks
# Verify venv exists
ls backend/.venv/bin/activate || echo "MISSING VENV"
# Verify node_modules exists
ls frontend/node_modules/.package-lock.json || echo "MISSING NODE_MODULES"
If either is missing, report the issue and STOP — do not attempt to install.
2. Run backend unit tests (Tier 1 — fast)
make test-unit
Expected: <120s. If tests fail, collect the failure output and handoff to Fix Test Failures.
3. Run backend integration tests (Tier 2 — Docker required)
Only if --run-integration is passed in $ARGUMENTS:
make test-integration
Expected: <600s. If Docker is not running or tests time out, report which integration tests passed/failed and continue with partial results.
4. Run frontend vitest tests (Tier 1 — fast)
make test-frontend
If tests fail, collect the failure output.
5. Run E2E tests (optional — requires running app)
Only if --e2e is passed in $ARGUMENTS:
make test-e2e
6. Generate coverage reports
make coverage
Review coverage percentages:
- Backend: check
backend/htmlcov/index.htmlor term report - Frontend: check
frontend/coverage/index.html
7. Linting gate
make lint
Output Format
## Full Test Suite Results
### Backend Unit Tests
- Total: N | Passed: N | Failed: N | Skipped: N
- Time: X.Xs
- [PASS/FAIL]
### Backend Integration Tests (if run)
- Total: N | Passed: N | Failed: N | Skipped: N
- Time: X.Xs
### Frontend Tests
- Total: N | Passed: N | Failed: N
- Time: X.Xs
### Coverage
- Backend: XX% (threshold: N/A)
- Frontend: XX% (threshold: 98% stmts, 95% funcs, 80% branches)
### Linting
- Backend (ruff): [PASS/FAIL]
- Frontend (eslint): [PASS/FAIL]
### Overall: [ALL_PASS / FAILURES_DETECTED]
Constraints
- NEVER run
pip installornpm install— report missing deps and stop. - If a test tier times out, report partial results rather than nothing.
- For integration tests: if Docker is not available, skip gracefully and note "Docker not available".
- Respect the anti-loop protocol: at attempt 3, re-check environment; at attempt 4, escalate.