context-engineer

You can bootstrap a structured Context repository by running a guided interview to generate a seeded three-layer directory model (drafts, workspace, context) with initial documents and synchronized indexes for AI use.

To keep AI context indexes synchronized, ingest raw inputs like notes and meetings so the system automatically routes content into the correct directory layer and updates tool-specific index files like CLAUDE.md or Cursor rules.

Yes, the context engineering workflow supports cross-tool compatibility by generating tool-specific index files, such as CLAUDE.md or Cursor rules, ensuring different AI assistants can access the current contextual truth.

You manage the document lifecycle by using maintain and evolve workflows that perform health checks and guide content transitions across the drafts, workspace, and context layers to keep the knowledge base accurate.

To bootstrap a structured context repository, you need a root SKILL.md frontmatter with a name and description, a three-layer directory model, and optional scripts, references, or assets folders to organize the knowledge base.