code-graph

Index and query a persistent code graph for symbol lookup and blast-radius analysis.

Updated Apr 21, 2026
One-click install
npx skills add https://github.com/joshuakessell/cornerstone --skill code-graph
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-graph
Source: https://github.com/joshuakessell/cornerstone/tree/main/.claude/skills/code-graph
Command: npx skills add https://github.com/joshuakessell/cornerstone --skill code-graph

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a persistent code graph to accelerate navigation, symbol lookup, and impact analysis in large codebases, reducing reliance on manual grep and file reads.

Core Features & Use Cases

  • Graph-first navigation: query codebase-memory-mcp for instant symbol lookup and function search.
  • Dependency and blast radius analysis: trace dependencies and understand change impact without scanning files.
  • Multi-language support: works across languages via MCP-backed graph.
  • Use case: onboarding, planning, and code exploration.

Quick Start

Query the graph to locate a symbol using search_graph before reading the full source file.

Frequently Asked Questions about code-graph

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze blast radius and trace dependencies in a large codebase without scanning files?

Blast-radius analysis is performed by indexing a persistent code graph and using the trace_call_path tool to instantly trace dependencies and understand change impact without manual file scanning.

What is the best way to do sub-ms symbol lookup across multi-language repositories?

Sub-ms symbol lookup is achieved by querying a persistent MCP-backed code graph using the search_graph tool, enabling instant function search and navigation across large multi-language repositories.

Do I need a specific MCP server to enable code graph navigation and impact assessment?

Yes, code graph navigation requires the codegraph MCP server to be accessible, providing the persistent graph index needed for sub-ms symbol lookup and impact assessment.

How does code graph navigation compare to manual grep for feature planning and onboarding?

Code graph navigation replaces manual grep by providing instant symbol lookup and dependency tracing via a persistent graph, yielding faster architectural understanding for onboarding and feature planning.

Can I get a code snippet directly from the graph after locating a function?

Yes, after using search_graph to locate a symbol, you can use the get_code_snippet tool to retrieve the relevant code snippet directly from the indexed graph.

When should I not use a persistent code graph for codebase exploration?

A persistent code graph may not suit tiny repositories or single-file edits where manual grep is faster, as its value comes from indexing large multi-language repositories for complex impact assessment.