research-cache

Cache cross-block file discovery results by writing a structured cache file at ON_EXIT and reading it at ON_ENTER before the next phase. This prevents re-discovering files already found in earlier phases, reducing duplicated discovery work.

Prevent re-discovering files during multi-phase root depth drills by storing read-verified files and Sibling DNA patterns in a structured cache. Subsequent phases read this cache to skip files already found and reuse exact patterns.

Sibling DNA pattern reuse works by storing exact patterns in a dedicated cache pool during initial discovery. Downstream phases read this pool to reuse the patterns directly, eliminating the need to re-identify matching structures.

The cross-block discovery cache writes its structured data to a markdown file located at /tmp/sti/<ticket-id>/workspace/research-cache.md. This file contains sections for Discovered Files, Sibling DNA Pool, Exploration Anchor, and Cache Status.

Read the discovery cache at ON_ENTER before phase PO.3.0 to skip files already found. This allows downstream phases to reuse read-verified results from earlier phases instead of duplicating the discovery effort.

A limitation of caching file discovery results is that only files read-verified in Phases 1-2 are cached. Files discovered but not verified will not be included, requiring re-discovery in later phases.