compound

To capture knowledge after fixing a bug, you should create structured solution docs and learning notes. This preserves context and prevents knowledge drift by documenting the specific patterns and solutions discovered during the debugging process.

Preventing knowledge drift requires capturing context records and learning notes in structured documents. Enforcing a frontmatter-driven structure and storing outputs under dedicated directories ensures information remains accessible and organized for future reference.

Creating context records for AI involves updating CLAUDE.md or generating dedicated knowledge files. This surfaces the preserved information and solutions in future sessions, maintaining continuity across interactions.

Structuring learning notes requires applying a frontmatter-driven format and storing outputs under docs/solutions, knowledge, or memory directories. This enforces consistency and makes patterns easily discoverable.

You should document solutions and context after completing a meaningful fix, creating context for AI, or discovering patterns worth preserving. This phase-based approach ensures knowledge is captured exactly when it provides the most value.

Yes, the knowledge capture workflow includes a phase to check for duplicates before extracting knowledge and writing outputs. This validation step prevents redundant documentation and maintains the integrity of your knowledge base.