You are an AI assistant helping me understand and improve a codebase.
Use the attached/generated files as the authoritative context.

we are in project path: .

Files for analysis:
- project/analysis.toon  (Health diagnostics - complexity metrics, god modules, coupling issues, refactoring priorities) [19KB]
- project/context.md  (LLM narrative - architecture summary, key entry points, process flows, public API surface) [3KB]
- project/evolution.toon  (Refactoring queue - ranked actions by impact/effort, risks, metrics targets, history) [1KB]
- project/project.toon  (Project logic - compact module view from code2logic, file sizes, dependencies overview) [10KB]
- project/README.md  (Documentation - complete guide to all generated files, usage examples, interpretation) [9KB]

Task:
- Summarize the architecture and main flows.
- Identify the highest-risk areas and propose a refactoring plan.
- If you suggest changes, keep behavior backward compatible and provide concrete steps.
- Highlight critical functions (CC ≥ 10) and god modules from analysis.toon.
- Prioritize refactoring actions by impact/effort ratio from evolution.toon.
- Validate entry points and public API surface match the architecture described.

Focus Areas for Analysis:
1. **Code Health Analysis** - Review complexity metrics, god modules, coupling issues from analysis.toon
2. **Refactoring Priorities** - Examine ranked refactoring actions and risk assessment from evolution.toon
3. **Architecture Overview** - Understand main flows, entry points, and public API from context.md
4. **Project Structure** - Analyze module organization and dependencies from project.toon
5. **Structured Data** - Use machine-readable formats for automated analysis and metrics extraction
6. **Visual Flow** - Review control flow diagrams and call graphs for architectural insights

Analysis Strategy:
- Start with analysis.toon for health metrics, then evolution.toon for action priorities
- Reference analysis.yaml for precise metrics and programmatic data

Constraints:
- Prefer minimal, incremental changes.
- Maintain full backward compatibility.
- Base recommendations on concrete metrics from the provided files.
- If uncertain, ask clarifying questions.
