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/manuelsiuro/mcp-android-server-python/claude-mdgit clone --depth 1 https://github.com/manuelsiuro/mcp-android-server-pythonWhat 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 | $0.09974 | $0.09974 |
| Opus 5 | $0.04987 | $0.04987 |
| Sonnet 5 | $0.01995 | $0.01995 |
| Haiku 4.5 | $0.00997 | $0.00997 |
Grade A, and why
mcp-android-server-python 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 — 1,232 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 Overview
MCP Android Agent is an MCP (Model Context Protocol) server for automating Android devices using uiautomator2. It exposes Android device control functionality through MCP tools that can be consumed by AI agents like Claude Desktop, GitHub Copilot Chat, or VS Code agent mode.
Requirements: Python 3.10+ (tested on 3.10, 3.11, 3.12, 3.13)
Development Commands
Environment Setup
Quick Install (Automated):
# Run the installation script (handles everything)
./install.sh
Manual Install:
# Create and activate virtual environment
uv venv
source .venv/bin/activate # On Windows: .venv\Scripts\activate
# Install dependencies (using uv - preferred)
uv pip install -e .
# Or using traditional pip
pip install -r requirements.txt
ADB Configuration
If ADB is not in your system PATH:
- Copy the environment template:
cp .env.example .env
- Edit
.envand configure your Android SDK path:
# Recommended: Set ANDROID_HOME
ANDROID_HOME=/Users/username/Library/Android/sdk
# Alternative: Set direct ADB path
# ADB_PATH=/Users/username/Library/Android/sdk/platform-tools/adb
The server checks for ADB in this order:
ADB_PATHenvironment variable (direct path)ANDROID_HOME/platform-tools/adb- System PATH
Running the Server
Quick Start:
# MCP stdio mode (for Claude Desktop, VS Code)
./start.sh
# HTTP mode (for development/testing)
./start-http.sh
Manual Start:
# Activate virtual environment first
source .venv/bin/activate
# MCP stdio mode (for AI agent integration)
python3 server.py
# HTTP mode with uvicorn (for development/testing)
uvicorn server:app --factory --host 0.0.0.0 --port 8000
Testing
# Run all tests
pytest
# Run with coverage
pytest --cov
# Run specific test file
pytest tests/test_server.py
Code Quality
# Lint and format with ruff
ruff check .
ruff format .
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 · 1,232 lines · 9,974 tokens per session scan A fc25e31fcbc8
mcp-android-server-python CLAUDE.md is an instructions file published in the GitHub repository manuelsiuro/mcp-android-server-python (0 stars, last pushed 11mo ago), licensed MIT. It adds 9,974 tokens to every session, about $0.0499 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
rails-ai-bridge AGENTS.md
Instructions for igmarin/rails-ai-bridge, covering agents.md — rails-ai-bridge development guide, architecture, key design decisions, testing and conventions.
native-devtools-mcp AGENTS.md
AGENTS.md instructions for vectora-foundry/native-devtools-mcp, covering agent context: native-devtools-mcp, 🧠 core reasoning loop, 🗺️ capabilities matrix (strategy guide), 🛠️ tool definitions & schemas and 1. vision & perception (the "eyes").
hinge-auto AGENTS.md
Instructions for houseunlimited/hinge-auto, covering agent instructions, what this project is, hard constraints (read before doing anything), setup flow (walk the user through this in order) and phase 1 — environment.
simulang-js CLAUDE.md
Instructions for simular-ai/simulang-js, covering @simular-ai/simulang-js, where the api is documented, mental model, logging is on by default and live log viewer (optional).
Interceptor AGENTS.md
AGENTS.md instructions for Hacker-Valley-Media/Interceptor, covering interceptor agent manual, install modes, core rules, background first (browser + macos) and surface decision.
agent-ios GEMINI.md
Instructions for squaduplabs/agent-ios, a project described as: Drive the iOS Simulator from the CLI for AI coding agents: accessibility snapshots with element refs, taps, gestures, typing, and screenshots.