agent-evolution

You need to create a new skill when pattern recognition across completed modules reveals recurring processes. This meta-skill analyzes finished work against a decision matrix to determine if documentation should crystallize into a reusable skill or just an AGENTS.md update.

Extract reusable patterns by reviewing completed tasks for recurring logic and applying knowledge crystallization. This skill provides a structured Draft Skill template and governance checks to accelerate converting those recognized patterns into standardized documentation.

The best way to standardize process improvement is applying a decision matrix to evaluate completed work and generate structured skill definitions. This ensures consistent documentation updates and enforces governance checks across all recurring modules.

Yes, this skill creation process works entirely standalone without external dependencies. It uses an internal decision matrix and draft structure to guide knowledge crystallization directly from your existing project artifacts.

You can update governance files by running this analysis after complex tasks. It determines when an AGENTS.md update is sufficient versus when a full skill definition is required, ensuring your governance documentation reflects emerging patterns.

The limitation of using a decision matrix for skill creation is that it relies on analyzing already completed work. It cannot proactively predict future patterns and only enforces governance checks on crystallized knowledge from past module development.