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/codesteward/codesteward-graph/claude-mdgit clone --depth 1 https://github.com/Codesteward/codesteward-graphWhat 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 | $0.03787 | $0.03787 |
| Opus 5 | $0.01894 | $0.01894 |
| Sonnet 5 | $0.00757 | $0.00757 |
| Haiku 4.5 | $0.00379 | $0.00379 |
Grade A, and why
codesteward-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 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 — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — Codesteward MCP Graph Server
What This Is
This is the Phase 1 standalone repository of the Codesteward platform. It is a self-contained Model Context Protocol server that parses source repositories into a structural code graph and exposes that graph as queryable tools for AI agents.
It has no dependency on the broader Codesteward agent platform (no deepagents, no LangChain, no workflows, no policy engine, no PostgreSQL). For graph storage it supports three backends: Neo4j, JanusGraph (Apache 2.0), or GraphQLite (embedded SQLite — no server needed, ideal for local dev via uvx).
Architecture
codesteward/ # repo root (workspace root, not published)
├── pyproject.toml # uv workspace — members: packages/*
├── docker-compose.neo4j.yml # Neo4j stack
├── docker-compose.janusgraph.yml # JanusGraph stack (Apache 2.0 alternative)
├── Dockerfile.mcp # Multi-stage build (supports both backends)
├── .gitlab-ci.yml # CI: test → build → publish PyPI → release GitHub
├── CHANGELOG.md
├── packages/
│ ├── codesteward-graph/ # PyPI: codesteward-graph
│ │ ├── pyproject.toml
│ │ └── src/codesteward/
│ │ └── engine/
│ │ ├── graph_builder.py # Orchestrates parse + graph write (backend-agnostic)
│ │ ├── backends/ # Graph database backend abstraction
│ │ │ ├── base.py # GraphBackend ABC
│ │ │ ├── neo4j.py # Neo4j backend (Cypher queries)
│ │ │ ├── janusgraph.py # JanusGraph backend (Gremlin queries)
│ │ │ └── graphqlite.py # GraphQLite backend (Cypher over SQLite — embedded)
│ │ ├── tree_sitter_parser.py
│ │ └── parsers/ # Per-language tree-sitter parsers (14 languages)
│ │ ├── base.py # LanguageParser ABC + LexicalNode/GraphEdge models
│ │ ├── python.py # GUARDED_BY (decorators, Depends), PROTECTED_BY (APIRouter), taint sources
│ │ ├── typescript.py # GUARDED_BY (decorators), PROTECTED_BY (router.use), taint sources (Express/NestJS)
│ │ ├── java.py # GUARDED_BY (annotations), taint sources (Spring MVC/Jakarta EE)
│ │ ├── go.py # PROTECTED_BY (Gin/Echo group.Use), taint sources (net/http/Gin)
│ │ ├── rust.py # PROTECTED_BY (Actix scope().wrap()), taint sources (Actix-web/Axum extractors)
│ │ ├── php.py # PROTECTED_BY (Laravel Route::middleware->group), taint sources (superglobals/Laravel/Symfony/PSR-7)
│ │ ├── csharp.py # PROTECTED_BY (ASP.NET MapGroup().RequireAuthorization), taint sources ([FromQuery]/HttpRequest)
│ │ ├── kotlin.py # GUARDED_BY (Spring annotations), taint sources (Spring/Ktor/Http4k)
│ │ ├── scala.py # taint sources (Play/Akka HTTP)
│ │ ├── c.py # taint sources (CGI/Mongoose/libmicrohttpd)
│ │ ├── cpp.py # taint sources (Crow/Drogon/Pistache/Oat++)
│ │ ├── cobol.py # regex-based (no tree-sitter grammar)
│ │ └── _ast_utils.py # Shared tree-sitter helpers + TreeSitterBase
│ └── codesteward-mcp/ # PyPI: codesteward-mcp
│ ├── pyproject.toml # depends on codesteward-graph
│ └── src/codesteward/
│ └── mcp/
│ ├── config.py # McpConfig (pydantic-settings) + load_config(); auto-detects backend
│ ├── server.py # FastMCP server, 5 tools (4 core + optional taint), --transport http|stdio
│ ├── setup.py # `codesteward-mcp setup` — auto-detect tools, register globally
│ └── tools/
│ ├── graph.py # Tool implementations (plain async fns)
│ └── taint.py # taint_analysis tool (invokes codesteward-taint binary)
└── tests/ # All tests run from workspace root
├── conftest.py # cfg / cfg_with_neo4j / cfg_with_janusgraph fixtures
├── test_engine/ # GraphBuilder + TreeSitterParser tests
│ ├── test_graph_builder.py
│ ├── test_tree_sitter_parser.py
│ └── test_taint_sources.py # Taint-source emission tests for all parsers
└── test_mcp/
└── test_graph_tools.py # Tests for all 4 MCP tool functions
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 · 263 lines · 3,787 tokens per session scan A 3a39d6f135f4
codesteward-graph CLAUDE.md is an instructions file published in the GitHub repository Codesteward/codesteward-graph (2 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 3,787 tokens to every session, about $0.0189 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 2026-08-31.
Other instructions, from other repositories
autotask-mcp dev_workflow.instructions.md
Guide for using Taskmaster to manage task-driven development workflows.
ai-workflow AGENTS.md
Instructions for cunhaax/ai-workflow, covering ai workflow template, rules — non-negotiable, project overview, commands and architecture.
trackly-cli CLAUDE.md
Claude Code instructions for trackly-app/trackly-cli, covering trackly-cli, tech stack, backend production source of truth, directory structure and key commands.
foggy-data-mcp-bridge-python CLAUDE.md
Instructions for foggy-projects/foggy-data-mcp-bridge-python, covering foggy data mcp bridge — python, 快速启动, 安装依赖, 运行测试 and 启动 mcp 服务（连接 docker mysql）.
McpServer AGENTS.md
AGENTS.md instructions for sharpninja/McpServer, covering agent instructions, session start, rules, byrd test gate and where things live.
AmbyKit CLAUDE.md
Instructions for ambystechcom/AmbyKit, covering claude code — notes for the ambykit repo and claude-specific.