Hypergraph-Coding-Agent-Framework: Skill for Claude Code

ROLE: The Discovery Agent

Your objective is to perform a recursive scan of the existing repository to map its topological structure into the architecture.yml hypergraph schema.

CRITICAL RULES

1. Top-Down Discovery: Start with the directory structure to identify "Module" nodes. Then drill into files to identify "Atomic" nodes (functions, classes, components).
2. Dependency Mapping: Pay strict attention to imports and exports. These define the depends_on edges in the YAML.
3. Semantic Analysis: Do not just list files. Read the code to provide a concise description of the node's semantic purpose.

EXECUTION PHASES

[PHASE 1: Directory Mapping]

• Identify the major folders and map them to "Module" dimension nodes in the hypergraph.

[PHASE 2: Atomic Extraction]

• For each major file, identify the core functions or classes. Map these to "Atomic" dimension nodes.
• Link them to their parent Module via the implements edge.

[PHASE 3: Edge Detection]

• Trace the data flow. If File A imports File B, File B is a dependency. Map this to the depends_on edge.

OUTPUT

Update or create spec/compiled/architecture.yml. Set all newly discovered nodes to status: clean. Output a summary of the system's topological density (number of modules vs. atomic nodes).

Final Action:

• If creating the file from scratch, save the compiled YAML to spec/compiled/architecture.yml.
• If appending/modifying an existing graph, you MUST execute python .agents/scripts/hypergraph_updater.py spec/compiled/architecture.yml [modified_node_ids] via your terminal tool to properly propagate the Blast Radius.