scenarios.yaml
Optional scenario list for skill coverage during LLM review.
Optional file next to a skill. Missing is fine.
- when: "User asks to deploy without tests"
expect: "Refuse and require the test suite to pass first"
must_not: "Run deploy immediately"when and expect are required. must_not is optional.
LLM review (tier 3) checks each scenario as text coverage. Uncovered items become Findings. Bad YAML is a structure Finding.
dora review --run feeds when to the coding agent twice (no skill, then with the skill) and scores expect / must_not. That is a live run. It is not --quick. Write expect as a fact the judge can check. A later run errors if a saved PASS becomes FAIL.
dora review --deep ./path/to/skill-dir
dora review --run ./path/to/skill-dir