graph: Instructions file for Claude Code

CLAUDE.md

graph CLAUDE.md is an instructions file for Claude Code from Graph-tl/graph. It costs 1,011 tokens per session, scanned A, original, MIT.

A coding guide for Graph, an MCP server that gives AI agents memory between work sessions. It stores project plans, dependencies, task ownership, evidence, handoffs, history, and knowledge in SQLite, a file-based database.

In plain words
What is it for?
Use it when changing Graph’s server, task-management tools, database layer, tests, or the automatic updates to agent configuration files.
Why use it?
It helps agents resume work with the project context instead of reconstructing plans and decisions from scratch each session.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: mentions CLAUDE.md; mentions Claude Code.

This is Graph-tl/graph's own configuration. It tells Claude Code how to work on graph itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything graph configures →

Reuse

Borrowing it

Nothing to install: this file belongs to Graph-tl/graph. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/Graph-tl/graph/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/Graph-tl/graph

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for graph CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/graph-tl/graph/claude-md/github.svg)](https://agentmods.dev/instructions/graph-tl/graph/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/graph-tl/graph/claude-md"><img src="https://agentmods.dev/badge/instructions/graph-tl/graph/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for graph CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/graph-tl/graph/claude-md"><img src="https://agentmods.dev/badge/instructions/graph-tl/graph/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 1,011 This file is loaded in full into every session.
When invoked 1,011 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.01011 $0.01011
Opus 5.5 $0.00404 $0.00404
Sonnet 5.5 $0.00202 $0.00202
Haiku 4.5 $0.00101 $0.00101

Measured 2mo ago against content hash d0172aeb028c, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

graph 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 2mo 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.

CLAUDE.md · 55 lines

How it starts

The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project

Graph is an MCP server that gives AI agents persistent memory across sessions. Agents decompose work into dependency trees, claim tasks, record evidence, and hand off to the next session automatically. Published as @graph-tl/graph on npm.

Commands

npm test                    # Run all tests (vitest, ~161 tests across 6 files)
npm run build               # Build with tsup (ESM, node20 target)
npx vitest run test/tools.test.ts  # Run a single test file
npx vitest run -t "creates a new project"  # Run a single test by name

Architecture

MCP server (src/server.ts) — registers 14 tools, handles tool calls via switch statement, manages version banner and agent file auto-update on first call.

Tool handlers (src/tools/*.ts) — one file per tool. Each exports a handleXxx(input, agent) function. Tools: open, plan, next, context, update, connect, query, restructure, history, onboard, tree, status, knowledge (4 sub-tools), agent-config.

Data layer:

  • db.ts — SQLite via better-sqlite3, WAL mode, schema migrations. DB is lazy-initialized on first tool call.
  • nodes.ts — CRUD for nodes (tasks). Handles depth computation, parent-child, project summaries. rowToNode() converts SQLite integers/JSON strings to typed objects.
  • edges.ts — Dependency edges with cycle detection (DFS). findNewlyActionable() walks the graph to find nodes unblocked by a resolution.
  • events.ts — Audit trail. Every mutation logs a timestamped event with field-level changes.
  • types.ts — Core interfaces: Node, Evidence, Edge, Event, NodeRow (SQLite row shape).

Validation (validate.ts) — Two error types: ValidationError (bad input) and EngineError(code, message) (business logic, e.g. cycle_detected, evidence_required, delete_protected).

Gating (gates.ts, license.ts) — License tier enforcement (free/pro). Gates checked in server.ts before calling handlers.

Read the full file on GitHub · 55 lines

Changes

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.

  1. 2mo ago First seen · 55 lines · 1,011 tokens per session scan A d0172aeb028c

Subscribe to this mod's changes

graph CLAUDE.md is an instructions file published in the GitHub repository Graph-tl/graph (7 stars, last pushed 8mo ago), licensed MIT. It adds 1,011 tokens to every session, about $0.0040 per session on Opus 5.5. 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 2026-08-31.

Related

graph CLAUDE.md compared with instructions from other repositories

graph CLAUDE.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
graph CLAUDE.mdGraph-tl/graphInstructions file78mo agoMITA1,011Claude Code
vscode buildNext.instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA9,719GitHub Copilot
next.js AGENTS.mdvercel/next.jsInstructions file143kyesterdayMITA7,461Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchainInstructions file149kyesterdayMITA4,680Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA5,516GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA3,093GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harnessInstructions file274k2d agoMITA4,313Codex, OpenCode