skill-smell-checkerlisted

# Skill Smell Checker Audit all 26 smells from Hong, Imani, and Ahmed's *From Anatomy to Smells*. Treat the five static results as script-derived facts and the other 21 as evidence-backed semantic judgments. ## Workflow 1. Locate the target `SKILL.md`. For a repository-wide audit, discover files with `rg --files -g '**/SKILL.md'`. 2. Run the deterministic harness: ```bash python SKILL_DIR/scripts/check_static_smells.py path/to/SKILL.md --pretty ``` Replace `SKILL_DIR` with this skill's directory. Do not manually override its five results. Use `--strict` only when a nonzero exit status should fail CI. 3. Read [references/smell-catalog.md](references/smell-catalog.md) completely. 4. Review every semantic smell against the full target skill and any bundled `scripts/`, `references/`, or `assets/` needed to judge delegation, validation, templates, or utility-script availability. 5. For each semantic smell, record `present`, `absent`, or `uncertain`, cite concrete file evidence, and give confidence. Never infer a clean result from a missing section title alone. 6. Produce the report using the template below. Rank fixes by security, execution correctness, discoverability, then context efficiency. 7. If remediation was requested, make the smallest behavior-preserving edits, rerun the static harness, and repeat the semantic review for changed smells. Track multi-file audits as `pending`, `checked`, or `blocked`; do not silently drop files o