self-improvement

Log AI agent errors by appending command failures and unexpected behaviors to structured markdown files like ERRORS.md within a .learnings directory. This captures mistakes and correct approaches to prevent recurring issues and refine agent behavior over time.

Structured learning capture records user corrections, knowledge gaps, and discovered best practices in markdown files. It enables continuous improvement by systematically documenting insights across various AI coding environments to accelerate development.

Promote captured learnings to project memory by integrating valuable insights into project documentation files like CLAUDE.md or AGENTS.md. This updates documentation and makes AI agents more reliable by retaining operational best practices.

Track feature requests and knowledge gaps by documenting missing capabilities requested by users in structured markdown logs. This records operational failures and corrections, supporting continuous improvement across AI coding environments.

Extract reusable skills from logged AI mistakes by reviewing structured markdown logs of errors and best practices, then promoting those valuable insights into new skills. This prevents recurring mistakes and refines agent behavior.

Continuous improvement logging works without external dependencies by using structured markdown files to capture errors, learnings, and feature requests. It relies on appending details to files like ERRORS.md within a .learnings directory.