Validates that natural-language trigger phrases correctly resolve to the expected skill. Each .txt file in prompts/ contains a phrase that should match the skill named by the filename. The test runner feeds each prompt through the skill-matching logic and asserts the correct skill is selected, catching regressions in trigger phrase routing.

Prompt filenames must map to real skills/<name>/SKILL.md and skills-codex/<name>/SKILL.md artifacts. Do not add legacy aliases here; use the current skill name that users should invoke.

Running

bash tests/explicit-skill-requests/run-all.sh