work-logs-knowledge-log

Knowledge logging writes down insights about code structure, design intent, and implicit rules discovered during sessions. It stores these logs in a centralized knowledge base at ~/.claude/knowledge/${repo_name}/ with timestamped filenames, creating a persistent record that bridges gaps between your mental model and what the codebase actually expresses.

Log newly recognized repository patterns, design rationale, operational constraints, business logic prerequisites, and implicit rules. Capture gaps between AI inferences and your actual understanding to build team knowledge about why code is structured the way it is and what assumptions drive its architecture.

Knowledge logs are stored under ~/.claude/knowledge/${repo_name}/ with filenames using Asia/Tokyo timestamps in format YYYY-MM-DD_HH-MM followed by a description. This centralized location makes logs easy to find, reference, and share across team sessions.

Yes. By documenting session-specific insights about code structure and design decisions, knowledge logs create a reproducible reference that helps you and your team maintain consistent understanding of the codebase across multiple sessions and different development contexts.

No. Knowledge logging requires only a working date command and write access to ~/.claude/ directory. It has no external dependencies or components; logs are plain Markdown files organized by repository name and timestamp.