self-improvement

To enable continuous improvement, log errors and user corrections by appending the details to structured markdown files in a `.learnings/` directory. This captures debugging failures and feedback for subsequent analysis.

The best way to capture knowledge gaps and feature requests is using structured logging in markdown files. This records missing capabilities and discovered best practices to prevent future errors and document needed corrections.

Yes, you can promote recurring learnings to permanent project memory files like `CLAUDE.md` or `AGENTS.md`. This elevates significant corrections from logs into the agent's core knowledge base, improving overall effectiveness.

You can extract recurring errors by reviewing structured error logs and identifying patterns in command failures or exceptions. These recurring learnings are then extracted into new, reusable skills to prevent future mistakes.

No specific dependencies are required. The approach relies solely on structured logging to markdown files for knowledge management, making it compatible with any software engineering environment that supports basic file operations.

The primary limitation is that markdown files require manual or subsequent analysis to identify knowledge gaps and promote best practices. The feedback loop depends on reviewing these logs rather than automated memory updates.