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/newcovid/localhands/claude-mdgit clone --depth 1 https://github.com/newcovid/LocalHandsWrote 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/newcovid/localhands/claude-md)<a href="https://agentmods.dev/instructions/newcovid/localhands/claude-md"><img src="https://agentmods.dev/badge/instructions/newcovid/localhands/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 | $0.02027 | $0.02027 |
| Opus 5 | $0.01014 | $0.01014 |
| Sonnet 5 | $0.00405 | $0.00405 |
| Haiku 4.5 | $0.00203 | $0.00203 |
Grade A, and why
LocalHands CLAUDE.md scanned grade A with 2 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
mint a one-time ticket and return a ready-to-run `curl` command carrying the Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
group so `taskkill /T` can reach descendants — a plain `subprocess.run` timeout only kills the How it starts
The opening of the file, as written. The whole thing — 147 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.
What this is
LocalHands is a local MCP server: a cloud agent (Feishu Aily / Kimi K3 / any MCP client)
connects over a tunnel and drives this workstation through 23 tools. The reasoning is remote;
this process is the hands. Read README.md for the product-level rationale — this file covers
what you need to change the code.
Commands
The project virtualenv holds the pinned mcp 2.x; a bare python may have the wrong one.
.venv\Scripts\python.exe -m pip install -e ".[dev]"
.venv\Scripts\python.exe -m pytest tests/ -q # 222 tests, no daemon/network/tunnel
.venv\Scripts\python.exe -m pytest tests/test_tools.py::TestEditFile -q # one class
.venv\Scripts\python.exe -m pytest tests/test_tools.py -k "encrypted" -q # by name
.venv\Scripts\python.exe -m localhands --config config.yaml --check # validate config, don't serve
.venv\Scripts\python.exe -m localhands --config config.yaml # serve (+ tunnel if enabled)
.venv\Scripts\python.exe -m localhands --config config.yaml --no-tunnel -v
ruff check . # config lives in pyproject.toml; ruff is not a declared dev dep
scripts\restart.ps1 kills whatever holds the port, sweeps orphaned ngrok/cloudflared,
relaunches detached, then polls /health until it answers before reporting success. Use it
rather than restarting by hand — it is the only path that verifies the daemon actually serves.
Status and log tails land in var/logs/restart.txt.
config.yaml is gitignored (it carries the auth token and machine paths); config.example.yaml
is the documented template and the file to update when adding a config key.
Architecture
Request path
tunnel → RateLimitMiddleware → BearerAuthMiddleware → asgi_router
├─ /mcp → StreamableHTTPSessionManager
├─ /sse, /messages/ → SseServerTransport
├─ /download/<t>, /upload/<t> → TransferEndpoints
└─ everything else → Starlette (/health, /)
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 · 147 lines · 2,027 tokens per session scan A 877f21b9facf
LocalHands CLAUDE.md is an instructions file published in the GitHub repository newcovid/LocalHands (0 stars, last pushed 23d ago), licensed MIT. It adds 2,027 tokens to every session, about $0.0101 per session on Opus 5. A static security scan graded it A with 2 findings (makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
mcp-camoufox AGENTS.md
Instructions for RobithYusuf/mcp-camoufox, covering mcp-camoufox — agent guide, product and repository map, hard constraints, working defaults and architecture invariants.
unbrowser CLAUDE.md
Instructions for protostatis/unbrowser, covering unbrowser, why this exists, architecture (locked), stealth mode (native, on by default) and host functions exposed to js.
mcp-server-amazon CLAUDE.md
Instructions for rigwild/mcp-server-amazon, covering claude.md, development commands, install dependencies (use -d flag for puppeteer), build typescript to javascript and clean mock html files.
whatsapp-automation CLAUDE.md
Instructions for sumioshi/whatsapp-automation, covering vocabulário de acompanhamento (desambiguação), recipe do monitor de mensagem (sem zumbi), modo de envio por chat (confirmar vs autônomo), mídia na nuvem (midiapendente) — não é erro and grupo renomeado ≠ chat morto (slug é estável).
browser-mcp CLAUDE.md
Claude Code instructions for drisplabs/browser-mcp, covering claude.md, project overview, build and development commands, build and quality checks.
google-flow-mcp CLAUDE.md
Instructions for roshanarnav25-sloth/google-flow-mcp, covering notes for ai coding agents, what this is, hard rules, before you claim it works and layout.