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/browseroperator/browser-operator-core/claude-mdgit clone --depth 1 https://github.com/BrowserOperator/browser-operator-coreWhat 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 | $0.01984 | $0.01984 |
| Opus 5 | $0.00992 | $0.00992 |
| Sonnet 5 | $0.00397 | $0.00397 |
| Haiku 4.5 | $0.00198 | $0.00198 |
Grade A, and why
browser-operator-core 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 2d 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 — 207 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.
Project Overview
Browser Operator is an AI-native browser built on Chrome DevTools frontend. It adds a multi-agent AI framework to the DevTools panel, enabling intelligent automation and web interaction through specialized AI agents.
Build & Development Commands
Initial Setup
# Prerequisites: depot_tools in PATH (https://chromium.googlesource.com/chromium/tools/depot_tools.git)
gclient sync
npm install
cp .env.example .env # Configure API keys
Build
npm run build # Standard build (runs gn gen automatically)
npm run build -- --watch # Watch mode for development
npm run build -- -t Debug # Build to out/Debug instead of out/Default
# Fast build (skip type checking and bundling)
gn gen out/fast-build --args="devtools_skip_typecheck=true devtools_bundle=false"
npm run build -- -t fast-build
Running DevTools with Custom Build
# Terminal 1: Build with watch
npm run build -- --watch
# Terminal 2: Serve the built files
cd out/Default/gen/front_end && python3 -m http.server 9000
# Terminal 3: Launch Browser Operator with custom DevTools
/Applications/Browser\ Operator.app/Contents/MacOS/Browser\ Operator \
--disable-infobars \
--custom-devtools-frontend=http://localhost:9000/ \
--remote-debugging-port=9222
Testing
npm run test # Unit tests (Karma/Mocha)
npm run webtest # E2E tests (Puppeteer)
npm run debug-webtest -- --spec=path/to/test # Debug specific test
npm run lint # ESLint
Eval Runner (Agent Testing)
Recommended: Use the eval-runner-analyst agent to run evals and get detailed analysis:
# In Claude Code, use the Task tool with eval-runner-analyst agent:
"Run the action agent evals with cerebras gpt-oss-120b"
"Test action-agent-checkbox-001 and action-agent-form-001"
"Compare V0 and V1 action agents on iframe tests"
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.
- 2d ago First seen · 207 lines · 1,984 tokens per session scan A 7d4429b5fa94
browser-operator-core CLAUDE.md is an instructions file published in the GitHub repository BrowserOperator/browser-operator-core (501 stars, last pushed 5mo ago), licensed BSD-3-Clause. It adds 1,984 tokens to every session, about $0.0099 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
eCan.ai CLAUDE.md
Instructions for scszcoder/eCan.ai, covering claude.md, 0. 废弃目录说明, 1. think before coding, 2. simplicity first and 3. surgical changes.
idun-agent-platform CLAUDE.md
Claude Code instructions for Idun-Group/idun-agent-platform, covering claude.md, project overview, constitution, repository map and quick commands.
koog CLAUDE.md
Instructions for JetBrains/koog, a project described as: Koog is a JVM (Java and Kotlin) framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise…
koog GEMINI.md
Instructions for JetBrains/koog, a project described as: Koog is a JVM (Java and Kotlin) framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise…
hindsight CLAUDE.md
Instructions for vectorize-io/hindsight, covering claude.md, project overview, development commands, local development (api + ui) and start both api server and control plane ui.
codedb AGENTS.md
Instructions for justrach/codedb, covering codedb agent guidelines, what codedb is (and isn't), review guidelines, pre-merge verification and security-sensitive areas.