recursive-training

Extract experiential memory from completed recursive-mode runs by running the recursive-training skill, which automates the consolidation of learnings from markdown artifacts into a canonical memory plane for future operations.

Experiential memory extraction is the process of consolidating durable learnings from completed runs. You need it when a repository has completed multiple recursive-mode runs and requires informed decision-making for subsequent operations.

Yes, Python is required to run the scripts that process markdown artifacts and extract memory items. The scripts handle the extraction and refresh logic to update the canonical memory plane post-Phase 8.

Yes, the skill provides startup guidance for recursive-mode runs without mutating the memory plane. It separates the extraction of learnings from the memory refresh to ensure existing memory remains intact during initial guidance.

The best way to consolidate cross-run learnings is to run the recursive-training skill after completing Phase 8, which extracts durable experiential memory from markdown artifacts and refreshes the canonical memory plane for subsequent runs.