Borrowing it
Nothing to install: this file belongs to sshh12/windows-computer-use-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/sshh12/windows-computer-use-mcp/main/CLAUDE.mdgit clone --depth 1 https://github.com/sshh12/windows-computer-use-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/sshh12/windows-computer-use-mcp/claude-md)<a href="https://agentmods.dev/instructions/sshh12/windows-computer-use-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/sshh12/windows-computer-use-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 5.1 | $0.02413 | $0.02413 |
| Opus 5 | $0.01207 | $0.01207 |
| Sonnet 5 | $0.00483 | $0.00483 |
| Haiku 4.5 | $0.00241 | $0.00241 |
Grade A, and why
windows-computer-use-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 6d 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 — 142 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
A Windows computer-use MCP server (Python + FastMCP, stdio) that gives an agent full control of the local Windows desktop: multi-monitor screen capture, native input injection, video recording, and a play-test loop. It runs ON the machine it controls (no resolution requests) and is built to work universally across native apps, games, and browsers. No security gating — full agent control by design.
Design principles (the decisions behind the surface — preserve these)
- Universal input is a hard requirement. Capture and input must work for native apps, games, AND browser/canvas games — no fullscreen workarounds, no per-app special-casing. If something works only for one class of app, it isn't done.
- The agent is the user; tokens are the scarce resource. Map tools to workflows (7 tools, not 35),
batch input to cut round-trips, and prefer non-image reads (
window get_text/ui_tree, OCR) over screenshots whenever text is enough. Return a single contact-sheet montage, never N frames. Downscale to 1568; write video/large output to a file, never inline. - Mirror the native
computer_20251124vocabulary (left_click,type,key,scroll,drag…) so the model is fluent on the first try. - Correctness beats cleverness in the capture/coordinate path.
capture_idmakes a stale-frame click a loud error, not a silent mis-click; onlyscreenshotsets the click frame; never steal keyboard focus to take a picture (PrintWindow captures in place). - Just-in-time feedback in tool results. When something looks wrong, the result should teach the agent the fix (which window received keyboard input; the browser-canvas hint; black-frame warnings) — don't rely on the static description alone.
- Read the real machine; never request a resolution. Multi-monitor + per-monitor DPI are first-class (the server controls the local desktop, unlike the sandboxed cloud tool).
- Validate against reality. Iterate via headless
claude -pagents running the latest code, and make tool-eval runs bail early and diagnose rather than flail.
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.
- 6d ago First seen · 142 lines · 2,413 tokens per session scan A 896b7ed74acc
windows-computer-use-mcp CLAUDE.md is an instructions file published in the GitHub repository sshh12/windows-computer-use-mcp (0 stars, last pushed 3mo ago), licensed MIT. It adds 2,413 tokens to every session, about $0.0121 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
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.