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/rushikeshmore/tor-mcp/claude-mdgit clone --depth 1 https://github.com/rushikeshmore/tor-mcpWrote 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.
[](https://agentmods.dev/instructions/rushikeshmore/tor-mcp/claude-md)<a href="https://agentmods.dev/instructions/rushikeshmore/tor-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/rushikeshmore/tor-mcp/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01132 | $0.01132 |
| Opus 5 | $0.00566 | $0.00566 |
| Sonnet 5 | $0.00226 | $0.00226 |
| Haiku 4.5 | $0.00113 | $0.00113 |
Grade A, and why
tor-mcp 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 4d 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 — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
tor-mcp
MCP server that routes LLM tool web requests through Tor with selective per-URL routing. Built on FastMCP, httpx[socks], stem.
Project structure
tor-mcp/
├── src/tor_mcp/
│ ├── __init__.py # Version
│ ├── __main__.py # `python -m tor_mcp` entry
│ ├── server.py # FastMCP server + 4 tools, validators, error handling
│ ├── tor.py # TorManager: SOCKS5h client + stem control
│ └── router.py # URLRouter: 7-priority pattern-based routing
├── tests/ # 89 tests, fully mocked
├── .github/workflows/ # CI: pytest + ruff on push/PR
├── pyproject.toml # Python 3.11+, hatchling build, ruff config
└── README.md # User-facing docs
Stack
- Python 3.11+
- FastMCP (mcp SDK, stdio transport)
- httpx[socks] for SOCKS5h Tor routing (DNS-safe)
- stem for Tor control protocol (circuit rotation)
- Pydantic v2 for input validation
Common commands
uv sync --dev # install deps
uv run python -m pytest # tests (mocked, no Tor needed)
uv run python -m ruff check src/ tests/ # lint
uv run tor-mcp # start MCP server (stdio)
uv run python -m tor_mcp # alternative launch
uv build # wheel + sdist
MCP tools
All tool names use the tor_ service prefix per MCP builder conventions.
| Tool | Annotations | Purpose |
|---|---|---|
tor_private_fetch |
read-only | Fetch URL with auto Tor/direct routing. force_tor and force_direct are mutually exclusive. |
tor_check_anonymity |
read-only | Verify Tor connection. Direct-IP compare with ipify → httpbin → ifconfig.co fallback chain. |
tor_new_identity |
side-effect | Rotate circuit. Requires control port. 10s rate limit enforced by Tor. |
tor_privacy_status |
read-only | Full health: connection, exit IP, control port, routing rules. |
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.
- 4d ago First seen · 115 lines · 1,132 tokens per session scan A a355fddba33f
tor-mcp CLAUDE.md is an instructions file published in the GitHub repository rushikeshmore/tor-mcp (2 stars, last pushed 3mo ago), licensed MIT. It adds 1,132 tokens to every session, about $0.0057 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
brosh CLAUDE.md
Instructions for twardoch/brosh, covering === user instructions ===, commands, installation, install the package in development mode and or install from pypi (when published).
axon AGENTS.md
AGENTS.md instructions for dinglebear-ai/axon, a project described as: Self-hosted RAG stack for crawl, scrape, search, ingest, query, and ask workflows with Qdrant, TEI embeddings, Chrome rendering, MCP/CLI/REST, and Gemini synthesis.
axon GEMINI.md
Gemini CLI instructions for dinglebear-ai/axon, a project described as: Self-hosted RAG stack for crawl, scrape, search, ingest, query, and ask workflows with Qdrant, TEI embeddings, Chrome rendering, MCP/CLI/REST, and Gemini synthesis.
chromectx AGENTS.md
Instructions for dyingg/chromectx, covering micro-commits, project shape, source layout, default store and adding a cli command.
chromectx CLAUDE.md
Instructions for dyingg/chromectx, a project described as: Search all your open tabs by title or full page content, and expose everything as RAG context for AI tools via MCP.
browser-control AGENTS.md
Instructions for keon/browser-control: Use this project as a tiny browser pipe, not an agent framework.