self-improvement

Capture command failures by logging the error message and execution context into structured markdown files such as `.learnings/ERRORS.md`. This records unexpected behaviors and exceptions to prevent recurring mistakes in AI agent workflows.

Agent error logging records command failures, user corrections, and knowledge gaps into markdown files. It integrates with agent hooks to automatically detect and log errors, facilitating continuous improvement and skill extraction for AI agents.

Yes, you can promote logged learnings to project memory files like CLAUDE.md or AGENTS.md. This transfers resolved corrections and discovered best practices from local error logs into persistent project memory or extracts them as new skills.

Track feature requests and user corrections by documenting requested capabilities that do not exist and logging user corrections into structured markdown files. This supports knowledge management by maintaining a record of knowledge gaps and recurring patterns.

Yes, automated error detection works with agent hooks for logging exceptions. The mechanism integrates with agent hooks to automatically detect errors and learning opportunities, capturing them into markdown files for agent refinement.

The best way to prevent recurring mistakes in AI agent workflows is to capture learnings, errors, and corrections into structured markdown files. This enables continuous improvement by documenting corrections and recurring patterns for agent refinement.