Graphify

Project analysis and graph extraction using Graphify

Skill metadata

Source	Bundled (installed by default)
Path	skills/software-development/graphify
Platforms	cli, telegram, discord

Reference: full SKILL.md

info

The following is the complete skill definition that Alice loads when this skill is triggered. This is what the agent sees as instructions when the skill is active.

Graphify Skill

Graphify is a powerful tool to extract, analyze, and visualize a codebase as a knowledge graph. It parses the codebase and generates graph.json and a compact summary in GRAPH_REPORT.md.

Instructions

Whenever you need to understand the big picture of a codebase, track complex dependencies, or find central modules (god-nodes), you should run Graphify.

1. Extract Graph: Run alice graphify (or /graphify in the CLI) in the root of the workspace. This is required instead of running uv run graphify directly, because the alice wrapper automatically injects your configured LLM API keys and backend settings into the extraction pipeline.
   • For a deeper semantic extraction, use the --deep flag: alice graphify --deep (or /graphify --deep).
2. Review Context: Once generated, Alice automatically reads GRAPH_REPORT.md on the next turn. You don't need to manually read it if it's already generated, but you can read GRAPH_REPORT.md if you need to refresh your memory.
3. Graph Data: If you need exact relationship mappings, read graph.json to find edge connections.

Useful Queries & Workflows

• Dependency Tracing: Search graph.json for specific modules to see what they depend on or what depends on them.
• God-Nodes: Look for nodes in GRAPH_REPORT.md or graph.json with exceptionally high edge counts.
• Incremental Updates: If the codebase changes significantly, suggest to the user to run alice graphify again to regenerate the graph.

Periodic Generation

If you have been making significant architectural changes, proactively ask the user: "Should we run /graphify again to update our project map?"