Skip to content

ldd-explain.md

Latest commit

 

History

History
21 lines (15 loc) · 1.51 KB

ldd-explain.md

File metadata and controls

21 lines (15 loc) · 1.51 KB
Raw
description Explain why LDD just invoked the skill(s) it invoked — cite the trigger phrase from the user's message that matched the dispatch table.

For the most-recently-invoked LDD skill (or skills) in this session, produce:

  1. Which skill fired — the full name.
  2. What triggered it — quote the exact substring of the user's message (or the earlier code context) that matched. Cite the row from the trigger-phrase table in skills/using-ldd/SKILL.md §"Trigger phrases".
  3. Why this skill and not another — if two dispatch rows could have matched (e.g. "bug" and "flaky test"), name the disambiguation rule applied.
  4. What the skill contributed — one line on the actual discipline (new Red Flag caught, refused fix pattern, rubric item enforced).
  5. Which rule you would have broken without it — the closest bad-faith alternative fix and why it would have been wrong.

No meta-theorizing. Concrete references to the user's message and the skills/using-ldd/SKILL.md table. 10 lines max.

If the user is asking about an auto-trigger that did NOT fire (they expected a skill and it didn't activate), check:

  • Was any trigger phrase present in their message?
  • Would the LDD: prefix have forced it?
  • Is the skill body's description tight enough to auto-match?

Report honestly. If the skill should have fired and didn't, that is a method-evolution signal — suggest the user file an issue with the .github/ISSUE_TEMPLATE/skill-failure.md template.