What problem does it solve? After a task succeeds, valuable lessons learned during execution are often lost. This Skill captures that experience from session logs and daily memory, then turns it into concrete, reviewable improvements for existing skill files. ## Core Features & Use Cases - Experience Extraction: Reads the session log and daily memory to identify what worked, new patterns, and edge cases solved. - Targeted Skill Location: Uses the skill-creator helper to scan workspace/skills/ and find the most relevant SKILL.md files to update. - Diff-Formatted Proposals: Produces clear diff blocks showing exactly which lines to change, with a summary of the new experience, without ever writing changes directly. - Use Case: After successfully completing a data-processing task that required an undocumented workaround, invoke this Skill to draft a diff adding that edge-case guidance to the relevant skill's instructions. ## Quick Start After confirming the task succeeded, ask the agent to run the skill-updater to extract lessons from this session and propose diff updates to the relevant SKILL.md files.
