understand-context

name	understand-context
description	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
argument-hint	[]

/understand-context

Reads .understand-anything/knowledge-graph.json (and domain-graph.json if present) and writes a structured architecture summary into the platform-appropriate context file. Run once after /understand — every subsequent agent session will have ambient architectural context without needing to query the graph.

Instructions

1. Check the graph exists.

   [ -f .understand-anything/knowledge-graph.json ] || { echo "No knowledge graph found. Run /understand first."; exit 1; }
   

2. Read the graph data. Use grep/jq to extract:

   • project section: name, description, languages, frameworks, analyzedAt, gitCommitHash (first 8 chars)
   • layers array: id, name, description, nodeIds length per layer
   • tour array: order, title, first sentence of description
   • File nodes with "complexity": "complex" — complexity hotspots
   • File nodes tagged "entry-point" — key entry points

3. Detect the target context file (check in this order):

   • [ -f CLAUDE.md ] → target is CLAUDE.md
   • [ -f AGENTS.md ] → target is AGENTS.md
   • [ -f .github/copilot-instructions.md ] → target is .github/copilot-instructions.md
   • None exist → target is AGENTS.md (create it)

4. Check for existing markers in the target file:

   • If <!-- understand-anything:start --> and <!-- understand-anything:end --> are both present: replace only the content between them (inclusive of the marker lines)
   • If absent: append the full block to the end of the file

5. Write the summary block with this exact format:

   <!-- understand-anything:start -->
   ## Codebase Architecture (understand-anything)
   
   **Project:** {name} — {description}
   **Languages:** {languages joined by ", "}   **Frameworks:** {frameworks joined by ", " or "none detected"}
   **Last analyzed:** {analyzedAt formatted as YYYY-MM-DD} (commit {first 8 chars of gitCommitHash})
   **Graph:** `.understand-anything/knowledge-graph.json`
   
   ### Architectural layers
   | Layer | Files | Purpose |
   |---|---|---|
   {one row per layer: | {layer.name} | {nodeIds.length} | {layer.description} |}
   
   ### Key entry points
   {For each file node tagged "entry-point": - `{filePath}` — {summary}}
   {If no entry-point tagged nodes: list the first 3 file nodes from the first tour step}
   
   ### Complexity hotspots (approach with care)
   {For each file node with complexity "complex": - `{filePath}` — {summary}}
   {If no complex nodes: write "No files currently marked complex."}
   
   ### Learning path ({tour.length} steps)
   {For each tour step: {order}. **{title}** — {first sentence of description}}
   <!-- understand-anything:end -->
   

6. If domain-graph.json exists, insert a "### Business Domains" section before the closing marker:

   ### Business Domains
   {For each node with type "domain": - **{name}**: {summary}}
   

7. Report to the user:

   ✓ Architecture summary written to {target file}
     Layers: {N}  Entry points: {N}  Hotspots: {N}  Tour steps: {N}
   
   Every agent session in this project will now start with codebase context.
   Re-run /understand-context after /understand to keep it fresh.