Browser MCP is a server that lets AI applications control a user's Chrome browser through the Model Context Protocol (MCP), a way for applications to call tools provided by another process. It runs on the user's computer and can work with signed-in sites, the user's browser profile, and open tabs. Its catalogue instruction relates to using this browser-control server from coding agents.
Borrowing it
Nothing to install: this file belongs to BrowserMCP/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/BrowserMCP/mcp/main/AGENTS.mdgit clone --depth 1 https://github.com/BrowserMCP/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/browsermcp/mcp/agents-md)<a href="https://agentmods.dev/instructions/browsermcp/mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/browsermcp/mcp/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/browsermcp/mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/browsermcp/mcp/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00861 | $0.00861 |
| Opus 5.5 | $0.00344 | $0.00344 |
| Sonnet 5.5 | $0.00172 | $0.00172 |
| Haiku 4.5 | $0.00086 | $0.00086 |
Grade A, and why
mcp AGENTS.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 9d 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 — 24 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Browser MCP's open-source core: the MCP server and the extension's browser control, which AI apps use to drive the user's own Chrome. See README.md for how the parts fit together.
packages/protocol(@browsermcp/protocol): the protocol between the extension and the daemon (frames, pairing, ports, close codes), and the extension IDs the daemon accepts.packages/mcp(@browsermcp/mcp): the MCP server, published to npm: the thin process each AI app runs, the daemon and its command engine. Read itsAGENTS.mdbefore changing it.packages/extension-core(@browsermcp/extension-core): browser control for a Chromium extension's background (addBrowserControl(),handleBrowserControlMessages()), a content script (@browsermcp/extension-core/content:handleFrameRemovalRequests(), which the extension declares in its manifest for<all_urls>in all frames, so it can remove other extensions' frames that block attaching, idle until the background asks) and popup (@browsermcp/extension-core/popup:useBrowserControl(),sendBrowserControlMessage()). It's source, not built: extensions bundle it.apps/extension(@browsermcp/extension): the extension people build from this repo, "Browser MCP (self-built)":extension-coreand a popup, in English. Its manifest's fixedkey(extension-key.ts) gives every build the IDSELF_BUILT_EXTENSION_ID, which the daemon accepts.
Conventions
- On its own: nothing here imports anything but these
@browsermcp/*packages and npm packages, and a fresh clone must install, typecheck, test and build. Refer to other packages by name (@browsermcp/mcp), not by path. - Protocol changes go into both sides at once: the daemon (
@browsermcp/mcp'ssrc/relay/) andextension-core. The protocol is versioned (RELAY_PROTOCOL_VERSION): each side tells the user which one to update when it meets an incompatible one. - Extensions built on
extension-corepass what they add (analytics, review prompts) as hooks (onEvent,onActivity), so browser control itself sends nothing anywhere. Its popup messages go through its ownruntime.onMessagelistener, beside any the extension has. - The popup's text: the smoke test finds the popup's elements by their English text ("Allow", "Code: …", "Connected…", "Stop", "Resume", "In use"), so keep those, or change the test with them.
- Chromium only: Firefox has no
debuggerAPI. - Tests: Vitest next to the code. Never bind or connect to ports 9009–9012 or touch
~/.browsermcp: use random ports,--portand a temporaryBROWSER_MCP_HOME, and stop only the processes a test started.
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.
- 9d ago First seen · 24 lines · 861 tokens per session scan A 321e714fe581
mcp AGENTS.md is an instructions file published in the GitHub repository BrowserMCP/mcp (7,297 stars, last pushed 3d ago), licensed Apache-2.0. It adds 861 tokens to every session, about $0.0034 per session on Opus 5.5. 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 2026-10-28.
mcp AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| mcp AGENTS.mdBrowserMCP/mcp | Instructions file | 7.3k | 3d ago | Apache-2.0 | A | 861 | Codex, OpenCode |
| plugin-files-modegriddynamics/rosetta | Instructions file | 370 | today | Apache-2.0 | A | 236 | GitHub Copilot |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |