Evaluate Graphify for skill dependency graph and cross-plugin analysis

[qte77]

Context

Graphify (16.5K stars, MIT) transforms codebases into queryable knowledge graphs via AST extraction (tree-sitter, 20+ languages) and LLM-powered semantic analysis.

CC integration surface

• Slash commands: /graphify ., /graphify query "...", /graphify path, /graphify explain
• PreToolUse hook: fires before Glob/Grep to surface graph structure
• CLAUDE.md injection: graphify claude install
• MCP server: python -m graphify.serve graph.json

Synergy with cc-utils-plugin

cc-utils-plugin has 18 plugins and 37 skills built from production workflows. At this scale, understanding cross-plugin relationships becomes non-trivial:

Specific use cases

1. Skill dependency graph: Which skills reference which tools, rules, or other skills? Graphify's semantic extraction can map these relationships beyond what grep catches
2. Cross-plugin analysis: Discover implicit dependencies between plugins (e.g., commit-helper and tdd-core may share patterns without explicit references)
3. Plugin onboarding: New users can /graphify query "which plugins help with Python development?" instead of reading all 18 plugin docs
4. Refactoring safety: Before reorganizing skills, query the graph for downstream consumers

Integration path

pip install graphifyy && graphify install   # adds hooks + CLAUDE.md
graphify .                                  # build graph of plugin repo
graphify query "which skills depend on tdd-core?"

No changes to cc-utils-plugin code — Graphify operates as an external analysis tool.

Goal

Use Graphify to maintain a queryable knowledge graph of the plugin ecosystem as it grows, making cross-plugin relationships visible and refactoring safer.

Sources

• Graphify: https://github.com/safishamsi/graphify
• Analysis: https://github.com/qte77/ai-agents-research/blob/main/docs/cc-community/CC-community-tooling-landscape.md