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/robithyusuf/mcp-stealth-chrome/agents-mdgit clone --depth 1 https://github.com/RobithYusuf/mcp-stealth-chromeWrote 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/robithyusuf/mcp-stealth-chrome/agents-md)<a href="https://agentmods.dev/instructions/robithyusuf/mcp-stealth-chrome/agents-md"><img src="https://agentmods.dev/badge/instructions/robithyusuf/mcp-stealth-chrome/agents-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 | $0.04144 | $0.04144 |
| Opus 5 | $0.02072 | $0.02072 |
| Sonnet 5 | $0.00829 | $0.00829 |
| Haiku 4.5 | $0.00414 | $0.00414 |
Grade A, and why
mcp-stealth-chrome AGENTS.md scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- `export_curl` emits both a `curl` string and an `http_request(impersonate='chrome')` call, and How it starts
The opening of the file, as written. The whole thing — 257 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mcp-stealth-chrome — Agent Guide
This guide applies to the whole repository and is the shared source of instructions for coding agents. It is not here to govern every decision — it exists to preserve the invariants that are easy to miss, almost all of which were learned from a shipped bug. Developer requests may override working defaults; hard constraints may only be overridden when the developer explicitly asks.
Product and repository map
An MCP server (stdio JSON-RPC) that gives an agent a stealth browser: 160 tools over
nodriver (direct CDP, no WebDriver), plus a
browserless HTTP path with a real Chrome TLS fingerprint. Published on PyPI as mcp-stealth-chrome.
uvx mcp-stealth-chrome → Python 3.11+ → FastMCP → nodriver → Chrome/Chromium (CDP)
└→ curl_cffi (Chrome JA3/JA4) → browserless HTTP
└→ opencv-python → template matching / vision
src/mcp_stealth_chrome/server.py— entry plus most tools, sectioned by══════banners.src/mcp_stealth_chrome/state.py—BrowserState(multi-instance), profile discovery, Chrome lookup, lock/PID helpers.src/mcp_stealth_chrome/snapshot.py— theSNAPSHOT_JSIIFE andformat_snapshot.src/mcp_stealth_chrome/helpers.py—ok/err, ref resolution, path confinement, JSON parsing.src/mcp_stealth_chrome/patches.py— runtime monkey-patches for nodriver/CDP drift.src/mcp_stealth_chrome/tools/—research.py,devtools.py,vision.py,network_http.py; each registers itself on import.tests/test_registry.py— the tool-surface gate.tests/test_reliability_e2e.pyandtests/test_agent_ergonomics.py— behaviour, real browser and fast unit respectively.README.md— the authoritative tool list.docs/— install, demo, research method, proof images.- Runtime data lives outside the repo:
~/.mcp-stealth/{profile,profiles,screenshots,exports, storage-states,identities,window-state}.
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 · 257 lines · 4,144 tokens per session scan A 6e709ac0a3eb
mcp-stealth-chrome AGENTS.md is an instructions file published in the GitHub repository RobithYusuf/mcp-stealth-chrome (23 stars, last pushed 17d ago), licensed MIT. It adds 4,144 tokens to every session, about $0.0207 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
cloakbrowser-mcp AGENTS.md
AGENTS.md instructions for swimmwatch/cloakbrowser-mcp, covering agents.md, about the project, golden rules, project layout and daily commands.
stealth-cli AGENTS.md
Instructions for Youhai020616/stealth-cli, covering stealth-cli, project structure, key architecture decisions, camoufox-js api (do not guess) and error handling.
browser-mcp AGENTS.md
AGENTS.md instructions for drisplabs/browser-mcp, covering snapshot & interaction pipeline — audit results, critical (agent can't complete tasks), major (agent works around it poorly), m2. batch vs. fallback coordinate space mismatch and m3. wrapping hidden inputs dropped from snapshot.
searchts CLAUDE.md
Claude Code instructions for capad-xyz/searchts, covering claude.md, project, commands, structure and conventions.
eyebrowse CLAUDE.md
Claude Code instructions for Evil-Bane/eyebrowse, covering claude.md — eyebrowse build notes, layout, run, pinned versions (lockstep matters) and key constraints (encoded in code — don't "simplify" away).
patchright-browser AGENTS.md
AGENTS.md instructions for rickicode/patchright-browser, covering agents.md — instructions for ai agents, what is this?, how to install, how to add mcp server and in your mcp config.