understand-anything

Analyzes a codebase's file structure, summaries, and import relationships to identify logical architectural layers and assign every file to exactly one layer.

Analyzes markdown files using pre-parsed structural data and LLM inference to extract knowledge graph nodes and edges (entities, claims, implicit relationships, topic clustering).

Reviews the output of merge-batch-graphs.py for semantic issues the script cannot catch. Recovers dropped nodes/edges and fills cross-batch gaps.

Analyzes codebases to extract business domain knowledge — domains, business flows, and process steps. Produces a domain-graph.json that maps how business logic flows through the code.

Analyzes batches of source files to produce knowledge graph nodes and edges. Extracts file structure, functions, classes, and relationships using a two-phase approach: structural extraction script followed by LLM semantic analysis.

Ask questions about a codebase using both the structural knowledge graph and the domain graph. Enhanced replacement for /understand-chat that includes business flow context.

Generate semantic embeddings for all knowledge graph nodes and save to .understand-anything/embeddings.json. Enables semantic search in the MCP server's search_nodes tool.

Check whether the knowledge graph is current relative to the git HEAD. Run before querying the graph to avoid reasoning from stale data.

Use when you need to ask questions about a codebase or understand code using a knowledge graph

Generate an architecture summary from the knowledge graph and inject it into CLAUDE.md, AGENTS.md, or .github/copilot-instructions.md so every agent session starts with codebase context

2 hooks across 2 events