| name | lewlm-validation |
| description | Plan and run validation for LewLM changes. Use for pytest selection, targeted test mapping, Python hook/script syntax checks, docs-only checks, long_running test decisions, and pre-PR confidence. |
| argument-hint | Describe changed files or behavior |
LewLM Validation
Use this skill to choose the smallest useful validation for a LewLM change.
Procedure
- Identify changed areas:
src\lewlm\api -> API route/schema tests.
src\lewlm\runtime or src\lewlm\routing -> runtime, scheduler, routing, and fallback tests.
src\lewlm\documents or src\lewlm\conversion -> document and conversion tests.
.github\agents, .github\skills, .github\instructions, .github\hooks -> customization validation and script syntax checks.
- Run the narrowest command first, for example:
python -m pytest tests\unit\test_registry.py -q
python -m pytest tests\integration\test_runtime_policy.py -q
python -m py_compile .github\hooks\scripts\safety_guard.py
- Use
python -m pytest -q for broad behavior changes or before PR handoff.
- Treat tests marked
long_running as opt-in unless the user asks for real-model benchmark coverage.
Output
Return:
- validation scope
- command
- result
- next command only if needed