enrich

/llm-wiki:enrich — Re-enrich Concept Docs

Re-runs enrichment on specific concepts (or all) using the okf-concept-enricher agent. Useful for refreshing stale docs after a source schema change.

Usage

/llm-wiki:enrich [concept-id…]

Without arguments, re-enriches ALL concepts in the bundle.

Steps

1. If concept IDs are specified, use them directly. Otherwise, walk the bundle and collect all concept IDs (from non-reserved .md files with a type field).

2. For each concept ID, read the existing doc:

   python3 <plugin_root>/scripts/okf_doc.py read <bundle_root> <concept_id>
   

3. Determine the source for fresh metadata:

   • If the concept has a resource URI pointing to BigQuery, use okf_bq.py describe.
   • For other concepts, re-read the raw source if available in raw/.
   • If no live source is available, enrich from the existing doc + context alone.

4. Fan out okf-concept-enricher subagents in parallel (one per concept). Each agent:

   • Reads existing doc
   • Gets fresh raw metadata
   • Writes augmented doc via okf_doc.py write
   • The PostToolUse hook validates each write

5. Report: N docs updated, M unchanged, any errors.

6. Suggest /llm-wiki:index to regenerate indexes after bulk updates.

Limiting scope

To re-enrich just one table:

/llm-wiki:enrich tables/events_

To re-enrich all reference docs:

/llm-wiki:enrich references/metrics/dau references/joins/events___users