ce-compound

Documenting a solved problem to build team knowledge is done by coordinating multiple subagents to extract symptoms, root causes, and solutions into a structured Markdown file. This creates a compounding knowledge base by placing the validated output under docs/solutions with cross-reference links.

Multi-agent documentation works by orchestrating Context Analyzer, Solution Extractor, and Related Docs Finder subagents. They process session history to capture symptoms, root causes, and prevention strategies, delivering a Markdown document that follows your project's YAML frontmatter schema.

Yes, you can generate both bug tracking reports and knowledge guidance docs. The workflow adapts to the specified problem_type, extracting either resolution details like root cause and prevention for bugs, or contextual guidance for general knowledge entries.

The best way to automate solution documentation with YAML frontmatter validation is using a multi-agent workflow that extracts problem details and generates Markdown. It automatically validates frontmatter fields against your repository schema and saves the final doc within the appropriate category.

You do not need existing docs to start, as the Related Docs Finder subagent searches your repository to automatically create cross-reference links. It connects your new solution documentation to existing related docs and issues to enrich the knowledge entry.

A limitation when using subagents to document workflows is that enriching learnings with session history depends on its availability. While the workflow integrates session context when present, documentation relies on the subagents accurately extracting symptoms and solutions from the provided data.