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/uiuing/browser-agent/agents-mdgit clone --depth 1 https://github.com/uiuing/browser-agentWhat 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 | $0.01806 | $0.01806 |
| Opus 5 | $0.00903 | $0.00903 |
| Sonnet 5 | $0.00361 | $0.00361 |
| Haiku 4.5 | $0.00181 | $0.00181 |
Grade A, and why
browser-agent 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 yesterday.
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 — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — guide for AI coding agents
Browser Agent is an open-source MV3 Chrome extension: the agent harness for the browser. Bring any model; the harness provides the agent loop, context assembly, tools, guardrails, memory and observability — and every claimed page effect is verified against page truth. TypeScript strict, pnpm monorepo, no backend.
Terminology: harness = the whole product. engine (src/engine/) = the page-action runtime (perception/grounding/execute/verify/failure handling), exposed to the loop as the page_act tool.
Setup & commands
pnpm install # workspace root; runs `wxt prepare` for the extension
pnpm build # fixtures + extension build, then typechecks bench/e2e
pnpm dev:ext # extension dev server (WXT hot reload)
pnpm --filter @browser-agent/extension compile # tsc --noEmit for the extension only
pnpm --filter @browser-agent/extension zip # release zip → packages/extension/.output/browser-agent-<ver>-chrome.zip
Extension build output: packages/extension/.output/chrome-mv3 (load unpacked in Chrome).
Tests — run these before claiming done
pnpm test:engine # fast, headless-DOM checks of the engine (perception/grounding/verify/heal)
pnpm test:e2e # full extension e2e with the scripted MockProvider; needs Playwright Chromium
pnpm shots # screenshot walkthrough of all UI surfaces → shots/
Notes for e2e:
- e2e builds the extension itself with
VITE_USE_MOCK=1(routes to a deterministic scripted mock — no API key needed). Never ship or commit a mock build;pnpm buildat root always produces the production build. - Real-model suites (
test:real,test:sitesinpackages/e2e) readREAL_BASE/REAL_KEY/REAL_MODELenv vars (never hardcode keys) and always rebuild in production mode;SKIP_BUILD=1reuses the previous build. - On Windows sandboxes, Playwright may need
PLAYWRIGHT_BROWSERS_PATHpointing at the localms-playwrightcache. - Chrome 137+ requires
--disable-features=DisableLoadExtensionCommandLineSwitchto load extensions from CLI (the e2e launcher already handles it).
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.
- yesterday First seen · 89 lines · 1,806 tokens per session scan A dd789a274df8
browser-agent AGENTS.md is an instructions file published in the GitHub repository uiuing/browser-agent (53 stars, last pushed 1mo ago), licensed MIT. It adds 1,806 tokens to every session, about $0.0090 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 2026-08-30.
Other instructions, from other repositories
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.
browser-bridge AGENTS.md
Instructions for koltyakov/browser-bridge, covering project guide, data flow, package map, where to find things and method dispatch (extension side).
codex-browser-bridge AGENTS.md
Instructions for DeliciousBuding/codex-browser-bridge, covering agents.md, 仓库级 skill, mcp 工具设计规范, 1. browser 层（src/browser.rs） and 2. mcp 层（src/mcp/ 目录）.
chromeboost GEMINI.md
Instructions for lordamdal/chromeboost, covering chromeboost — gemini instructions, what chromeboost is, when to use chromeboost (be proactive), hard rules — never break these and standard flow pattern.
chromeboost CLAUDE.md
Instructions for lordamdal/chromeboost, covering chromeboost — repo-developer guide, what chromeboost is, repository layout, development commands and tests/antibot/ (run locally; not in ci).
sidebutton AGENTS.md
Instructions for sidebutton/sidebutton, covering agents.md, project, build & run, code style and testing.