Summary
Auditing the 207-gist knowledge base revealed 6+ inconsistent naming conventions: bare filenames (54), essay-length descriptions (16),ctx: prefixes (8), skill trigger text (9), and ~100 medium prose descriptions. TOOLS.md defined a convention ([ORG]:[TYPE]:[SLUG].[ext]) but zero gists followed it.
What changed operationally
Defined canonical convention:{org}:{type}:{slug} — {summary} with 17 content types. Built a YAML manifest indexing all 207 gists with proposed canonical names, priority tiers (P0-P4), and org/type classification. The manifest enables programmatic lookup by org, type, or priority.
Business impact
- 5 SPEC gists and 5 data-payload gists (consumed by runtime) are now P0-indexed and findable
- 19 skill gists previously identifiable only by reading their full descriptions can now be found by
devarno:skill:{name} - Migration path defined:
gh gist edit {ID} --desc "{canonical}"— 207 commands to execute
Operational takeaway
Define gist naming conventions before the collection passes ~20 gists. Retrofitting costs 10x more effort than establishing the convention early. The convention must accommodate all content types — a narrow format like[ORG]:[TYPE].[ext] will be outgrown within a month.