Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/cognitx-leyton/codegraph/claude-mdgit clone --depth 1 https://github.com/cognitx-leyton/codegraphWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.02607 | $0.02607 |
| Opus 5 | $0.01303 | $0.01303 |
| Sonnet 5 | $0.00521 | $0.00521 |
| Haiku 4.5 | $0.00261 | $0.00261 |
Grade A, and why
codegraph CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 130 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Guidance for Claude Code (and similar coding agents) working on this repo.
Project summary
codegraph (package: codegraph) is a Python tool that indexes TypeScript and Python codebases into a Neo4j property graph. It walks source with tree-sitter, recognises framework constructs (NestJS controllers / injectables / modules, React components and hooks, TypeORM entities, GraphQL operations, Python classes and decorators), and loads typed nodes + edges into Neo4j. Downstream consumers:
- CLI:
codegraph index,codegraph query,codegraph validate,codegraph wipe— Typer app. - MCP server:
codegraph-mcpstdio server with 16 read-only tools. Optional extra (pip install "codegraph[mcp]"). - REPL: interactive Cypher shell at
codegraph repl.
This repo is itself Python, and codegraph parses Python since Stage 1 shipped. So we can dogfood: Claude Code can query the graph of codegraph-the-codebase while implementing codegraph-the-tool.
Using the graph during development
Two slash commands wire the local Neo4j graph into your Claude Code workflow. Use them — they exist to catch mistakes the language server can't.
/graph <cypher> — query the live graph
Read-only Cypher against bolt://localhost:7688. Prefer this over manual codegraph query in Bash because the slash command has the permissions pre-approved and a body of canonical query patterns. See .claude/commands/graph.md for examples.
When to run /graph:
- Before renaming a class, function, or file → check blast radius with
MATCH (f:File)-[r:IMPORTS_SYMBOL]->(g:File) WHERE r.symbol = 'X' RETURN f.path, g.path. - Before deleting a function or method → confirm nothing depends on it.
- Before moving a module → count incoming IMPORTS edges; you'll need to update every caller.
- When asked "who calls X?" or "where is Y used?" — almost always a one-query answer.
- When in doubt about what exists —
/graph "MATCH (c:Class) WHERE c.name CONTAINS 'Foo' RETURN c.name, c.file"is faster and more exhaustive than grep.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 130 lines · 2,607 tokens per session scan A cba197e193ac
codegraph CLAUDE.md is an instructions file published in the GitHub repository cognitx-leyton/codegraph (11 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 2,607 tokens to every session, about $0.0130 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
graphify AGENTS.md
AGENTS.md instructions for Graphify-Labs/graphify: This project has a graphify knowledge graph at graphify-out/.
ops-codegraph-tool CLAUDE.md
Claude Code instructions for optave/ops-codegraph-tool, covering claude.md, codegraph workflow, before modifying code, after modifying code and navigation.
arbor CLAUDE.md
Claude Code instructions for getArbor-dev/arbor, covering claude.md, commands, build, test all crates and test single crate.
navegador CLAUDE.md
Instructions for ConflictHQ/navegador, covering navegador — claude context, what it is, stack, package layout and falkordb connection.
aptu-coder AGENTS.md
AGENTS.md instructions for clouatre-labs/aptu-coder, covering agents.md, project structure, ci runners, commands and observability.
aptu-coder copilot-instructions.md
Copilot instructions for clouatre-labs/aptu-coder, covering copilot instructions, assigning issues to copilot, pr iteration, copilot code review and design references.