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/linvireo/browsertap-mcp/agents-mdgit clone --depth 1 https://github.com/LinVireo/browsertap-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/linvireo/browsertap-mcp/agents-md)<a href="https://agentmods.dev/instructions/linvireo/browsertap-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/linvireo/browsertap-mcp/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 | $0.05456 | $0.05456 |
| Opus 5 | $0.02728 | $0.02728 |
| Sonnet 5 | $0.01091 | $0.01091 |
| Haiku 4.5 | $0.00546 | $0.00546 |
Grade A, and why
browsertap-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 5d 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 — 394 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Notes for an AI agent working in this repository. Only things the code does not show on its own -- the traps that cost someone an afternoon. General coding style, the gate commands and the release flow live in CONTRIBUTING.md (简体中文).
What each tool does is not in this file. The full parameter table for all 55
tools is the ## Tools section of README.md
(简体中文). That table is the single authoritative list; do
not copy it here, it will go stale.
1. Three processes, three different ways a change takes effect
This is the first thing to check when an edit appears to do nothing.
| Process | What it is | When your edit takes effect |
|---|---|---|
| MCP server | one per client session, short-lived | immediately on an editable install |
| bridge daemon | pythonw -m browsertap_mcp.bridge, long-lived, outlives every server |
only after a restart |
| Chrome extension | unpacked MV3 extension loaded from src/browsertap_mcp/chrome_extension |
only after a manual reload |
The bridge runs its own BrowserBridge and keeps its own
default_session_id. A request with sessionId: None is therefore routed by the
bridge, not by the server. After changing any session-resolution logic in
browser_bridge.py, restart the bridge before testing, or you will conclude the
change had no effect:
browsertap bridge --restart # Chrome does not need restarting
The extension reconnects by itself within a few seconds; leave the browser alone.
Reloading the extension cannot be automated. Do not spend time on it:
chrome.runtime.reload()restarts the service worker but does not re-readbackground.jsfrom disk. You reconnect to the same old build.chrome.management.setEnabled(self, false)does force a re-read, but the only thing that could re-enable the extension is the extension itself, so the bridge goes down permanently and needs manual recovery. The extension now refuses it (self_disable_unsupported), because a rule that lives only in this file is one an agent reaching forset_extension_enablednever reads -- and that tool is the exposed path to exactly this call. Do not remove the guard to "force a reload"; there is no automated reload.
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.
- 5d ago First seen · 394 lines · 5,456 tokens per session scan A 0ae6b50d6f2c
browsertap-mcp AGENTS.md is an instructions file published in the GitHub repository LinVireo/browsertap-mcp (2 stars, last pushed 3d ago), licensed MIT. It adds 5,456 tokens to every session, about $0.0273 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
nekoro-browser AGENTS.md
Instructions for zeshuochen/nekoro-browser, covering code priorities, testing and contributing.
daedalus AGENTS.md
AGENTS.md instructions for Nitjsefnie-Harness-Commons/daedalus: Daedalus is a Chrome-extension-based remote control system for matching web pages — eval bridge, screenshots, CDP, cookies, network capture, hotfixes, declarativeNetRequest blocking.
daedalus CLAUDE.md
Claude Code instructions for Nitjsefnie-Harness-Commons/daedalus, a project described as: Browser bridge for agents: drive a real Chrome instance over HTTP, MCP or CLI.
rustwright AGENTS.md
AGENTS.md instructions for Skyvern-AI/rustwright, covering repository instructions, public boundary, pull-request review, local verification and browser launches.
opentabs AGENTS.md
Instructions for opentabs-dev/opentabs, covering project instructions for codex, project overview, architecture, you are connected to this project's mcp server and workflow skills.
chrome-agent CLAUDE.md
Claude Code instructions for sderosiaux/chrome-agent, covering chrome-agent v0.16.0, architecture, build & test, release and → requires npmtoken in github secrets.