chronicler

To automate lessons learned, this Skill scans recent git history to identify patterns and propagates those improvements directly into DEVELOPMENT_RULES.md, CLAUDE.md, and agent SKILL.md files. It generates a retrospective draft and summarizes actionable rule updates.

Updating agent skills based on project mistakes is handled by analyzing recent agent activity to extract lessons, then applying those changes across project-wide artifacts like all agent SKILL.md files. This maintains alignment and prevents repeating errors.

Retrospectives coordinate with scanner scripts by aligning checks with real-world findings from recent work. This supports process improvement by ensuring scanner scripts validate against actual patterns extracted during the retrospective analysis.

You can propagate documentation updates across multiple project files at once by applying extracted lessons to project-wide artifacts such as DEVELOPMENT_RULES.md, CLAUDE.md, and all agent SKILL.md files simultaneously with documented rationale and clear guardrails.

Analyzing recent agent activity to prevent repeating mistakes involves scanning recent work, gathering evidence, and performing pattern analysis. The extracted lessons are then propagated into rules and scanners to maintain project alignment.

Generating a retrospective draft does not require manual evidence gathering because the Skill runs a metrics-gathering flow that coordinates evidence collection, pattern analysis, and change propagation automatically with documented rationale.