## Summary
Two skill-forge-judge cycles on skill-forge-update surfaced seven targeted improvements which were applied via skill-forge-hitl. The skill moved from B (86%, 103/120) to A (91%, 109/120) on the U+S rubric. Changes: terminated two ambiguous decision branches, upgraded Phase 4 from a one-liner to a structured close-out with project-level default, added a Phase 3 zero-change guard, inlined trigger scenarios into the description, opened the drift check with a thinking-lens question, and added a sixth NEVER protecting hitl as Phase 2's exclusive apply mechanism.
## Acceptance Criteria
- [x] `skill-forge-update/SKILL.md` description ends with "Use when an existing SKILL.md needs revision, modification, or improvement."
- [x] `## NEVER` section contains six rules (verify with `grep -c "^- \*\*NEVER" skill-forge-update/SKILL.md` → 6)
- [x] Phase 4 close-out template contains a `**Judge:**` line
- [x] Re-running `/skill-forge-judge update` returns grade A (≥90%)
## Blockers
None
---
*Created with [gh-weld](https://github.com/WrathZA/github-weld)*
Forged skill-forge-update from B (86%, 103/120) to A (91%, 109/120) by applying seven judge findings across two cycles.
**What changed:**
- Terminated two ambiguous decision branches with explicit re-entry semantics
- Upgraded Phase 4 from a one-line file write to a structured close-out with project-level skills as the default and a symlink check that only fires for user-level installs
- Added a Phase 3 zero-change guard mirroring NEVER #5: surface `(a)/(r)/(q)` instead of silently advancing
- Inlined "Use when an existing SKILL.md needs revision..." into the description for harness-independent activation
- Opened the drift check with a thinking-lens question to shift from symptom-matching to first-principles evaluation
- Added NEVER #6 protecting hitl as Phase 2's exclusive apply mechanism
- Added a `**Judge:**` line to the close-out template to absorb the new Phase 3 reviewed-not-applied annotation
Shipped in PR #23.