Better the emsemioverse: $ARGUMENTS
Operation
Run python .claude/skills/better-emsemioverse/assess.py to produce a
gap report. The script checks the repository against five ASR requirements:
- Directory organization — directories with content but no index.md
- Skill manifests — SKILL.md files missing required manifest frontmatter
- Semantic frontmatter — files without typed relation fields
- Formal artifacts — files without colocated .ttl, .lean, .py, .json
- Linking — wikilinks remaining in published content
Use --check <name> to run one check. Use --json for structured output.
After assessment
Pick the highest-impact gap from the report. Close it using existing skills. If no skill covers the operation, create or extend one.
After closing the gap, run assess.py again to verify the count changed.
Then run /improve-skill better-emsemioverse if the assessment missed
something or the script needs a new check.
Extending
When a new ASR specification is written, add a corresponding check
function to assess.py. The check should:
- Reference the spec by path
- Produce a count of conforming vs non-conforming items
- List the non-conforming items
The script is the ground truth for “how close is this to a proper ASR.”