reflection

To optimize your AI agent ecosystem, run the Reflection skill to analyze conversation context and uncommitted git changes. It generates a supervisor analysis report identifying improvement opportunities for agents, skills, and commands based on your actual workflow.

Workflow automation reflection is a supervisor-first process that evaluates conversation patterns and repository state to propose improvements to the AI agent ecosystem. It identifies optimization opportunities before spawning builder agents to implement concrete changes.

Yes, you can update CLAUDE.md documentation based on git analysis. The skill synchronizes documentation to reflect the evolving agent ecosystem after proposing and implementing improvements to agents, skills, and commands.

Yes, agent ecosystem optimization always requires explicit user approval before making changes. The skill operates as supervisor first, producing an initial analysis report, then only spawns builder agents after you approve the proposed improvement plan.

The best way to coordinate builder agents is through a supervisor-first workflow that analyzes conversation context and git state, then spawns specialized builder agents for agent, skill, and command changes after user approval of the initial plan.