recall

To retrieve knowledge-base entities, scan all .evolve/entities/ directories across private and subscribed sources, parse YAML frontmatter and markdown bodies, then format provenance-annotated entity outputs for task context injection.

Searching markdown files with YAML frontmatter involves parsing the frontmatter fields like type, trigger, and rationale alongside the body content to evaluate usefulness and determine relevance to the current task execution.

Yes, retrieving task context includes scanning both private sources and subscribed clones, covering read- and write-scope files, to locate candidate entity files and preserve provenance through accurate annotations.

Building task context requires discovering entity files in .evolve/entities/ subdirectories, reading their YAML frontmatter and markdown bodies to assess relevance, and formatting concise context annotations for task planning.

Entity retrieval works across both private and subscribed knowledge sources simultaneously, searching all .evolve/entities/ files including read- and write-scope clones to surface relevant context-aware entities.

Entity retrieval is limited to files within .evolve/entities/ directories and requires valid YAML frontmatter and markdown structure to parse relevance, meaning improperly formatted or misplaced entity files will not be discovered.