graphify code lookup

The problem

Scanning large codebases to understand dependencies or call chains wastes tokens and time. Grepping files often misses indirect relationships and inferred architectural patterns that aren't obvious in raw text.

What it does

• Reads the GRAPH_REPORT.md to identify core "god nodes" and high-level community hubs in the architecture.
• Executes graphify explain to pull exact file locations, line numbers, and dependency degrees for specific components.
• Runs graphify path to trace the shortest call or dependency chain between two disparate parts of the system.
• Distinguishes between EXTRACTED facts (hard code edges) and INFERRED connections with confidence scores.
• Calculates the blast radius of proposed changes by reporting inbound edges to a component.

Frameworks & tools

Graphify, Linux/macOS, and existing graphify-out/ JSON graph data.

Why this beats prompting it yourself

Manually prompting an AI to "find how A calls B" requires feeding it dozens of files, which is expensive and prone to context window limits. This skill queries a pre-computed structural graph, providing grounded citations and line numbers without reading a single source file.

Use cases

• Tracing a request flow from a controller through middleware to the database.
• Identifying the most critical, highly-connected modules in a legacy codebase.
• Determining which services will break before refactoring a shared utility.
• Onboarding to a new project by visualizing community clusters and hubs.

Known limitations

Requires an existing Graphify graph; will not function if the graphify-out/ folder is missing. Stale graphs may report incorrect line numbers if the code has changed significantly since the last map.