yogsoth-ai Knowledge Compilation
Compile research findings into structured vault pages. Triggered after each strategy completes, or when CC judges accumulated information density warrants compilation.
Available SOPs
wiki-search— check for existing pages before creating duplicateswiki-ingest-source— write immutable source pageswiki-compile-page— create/update wiki pages with edgeswiki-add-edge— create typed relationships between pageswiki-lint-fix— run lint in report mode at end
Guiding Principles
- Assess before acting. Scan incoming material. Identify what's new vs what updates existing pages.
- Deduplication first. Always search before creating. Near-duplicates should be merged, not added.
- Edges are first-class. Every new page should have at least one edge connecting it to the existing graph. Orphans are failures.
- Source immutability. Source pages capture raw material verbatim. Wiki pages synthesize and evolve.
- Batch efficiency. Group related operations. Create all pages for a concept cluster, then wire edges, then index once.
Minimum Yield
Execution Guidance
CC has full autonomy over:
- Which SOPs to invoke and in what order
- Whether to spawn a subagent for complex ingest decisions
- How to partition material across pages
- When to create new pages vs update existing ones
Typical flow (reference, not prescription):
- Search vault for existing coverage of incoming topics
- Decide page operations: new sources, new wiki pages, updates, edges
- Execute operations via SOPs
- Run vault_lint (report mode) to verify no broken links introduced
Context-Management Integration
After compilation completes, trigger context-checkpoint to persist vault state changes to the context layer.