Browser4 is an AI-native browser engine that lets autonomous agents and people control real browsers, extract information from web pages, and automate tasks at scale. It combines browser interaction through a Rust CLI and MCP with deterministic selectors, machine-learning clustering, and language-model extraction. The catalogue entries provide skills and instructions for using Browser4.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/platonai/Browser4npx agentmods add skills/platonai/browser4/coworkerWrote 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/skills/platonai/browser4/coworker)<a href="https://agentmods.dev/skills/platonai/browser4/coworker"><img src="https://agentmods.dev/badge/skills/platonai/browser4/coworker/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/skills/platonai/browser4/coworker"><img src="https://agentmods.dev/badge/skills/platonai/browser4/coworker.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 4 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high YARA Match · line 4 YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).Fix: Remove offensive tool references and exploit code. Legitimate agent skills should not contain penetration testing tools, exploit frameworks, or reconnaissance utilities.
- medium Excessive Agency · line 93 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 107 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 163 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00050 | $0.04451 |
| Opus 5.5 | $0.00020 | $0.01780 |
| Sonnet 5.5 | $0.00010 | $0.00890 |
| Haiku 4.5 | $0.00005 | $0.00445 |
Grade A, and why
coworker 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 1mo 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 — 186 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Coworker
This repository contains a file-queue automation system called Coworker. The active implementation is the PowerShell worker in coworker/scripts/engineer.ps1: it watches task files, renames them, runs GitHub Copilot against the repository, logs the run, and routes the task through review/approval/push folders. (coworker/README.md:1-40, coworker/scripts/engineer.ps1:50-80, coworker/scripts/engineer.ps1:413-717)
What Coworker is for
- Use Coworker when you want an agentic workflow driven by task files in the repo, not by chat alone.
- A task can be a plain Markdown file or a structured file with
Title:,Description:, andPrompt:. If the structured header is missing, the full file becomes the prompt. (coworker/scripts/engineer.ps1:11-20,coworker/scripts/engineer.ps1:536-541) - The worker runs GitHub Copilot CLI with broad repo access to execute the task against the current repository. The helper command is configurable in
coworker/scripts/config.psd1. (coworker/scripts/workers/gh-copilot.ps1:37-65,coworker/scripts/config.psd1:1-12) - Important: the current/live workflow is file-based. The folder-based “story.md / analysis / plan / design / impl.patch / e2e” pipeline described in
coworker/docs/architect/orchestrator.mdandcoworker/scripts/architect/orchestrator.ps1exists as design/legacy material, but it is not the main coworker entrypoint used bycoworker.ps1. (coworker/docs/architect/orchestrator.md:1-38,coworker/scripts/architect/orchestrator.ps1:17-29,coworker/scripts/engineer.ps1:57-68)
Main workflows and scripts
1. Main task execution
- Primary entrypoint:
./coworker/scripts/engineer.ps1 - It ensures the task directories exist, optionally accepts a task file path, moves that file into
coworker/tasks/main/1ready, generates a descriptive kebab-case filename, moves it to2working, runs Copilot, writes logs, then moves the task tomain/3completeor5approved. (coworker/scripts/engineer.ps1:22-35,coworker/scripts/engineer.ps1:82-95,coworker/scripts/engineer.ps1:417-500,coworker/scripts/engineer.ps1:551-713) - The prompt given to Copilot explicitly says: finish the task described in the file, but do not move that task file yourself. The script handles routing after execution. (
coworker/scripts/engineer.ps1:531-545)
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.
- 1mo ago Changed 3cd24f24bef5
- 1mo ago First seen · 186 lines · 50 tokens per session scan A 066c4e835e42
coworker is a skill published in the GitHub repository platonai/Browser4 (1,164 stars, last pushed today), licensed Apache-2.0. It adds 50 tokens to every session and 4,451 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
eric-task-master
Run durable local Chrome automation with the installed Task Master CLI. Use when an Agent needs to create and execute a one-off browser task, follow its progress, or manage its Chrome Profile.
worker-integration
Worker-Agent integration for intelligent task dispatch and performance tracking.
agentmail
Use your assigned AgentMail inbox to read email tasks, explicitly send or reply, and check delivery. Provided automatically by your inbox assignment.
jetson-diagnostic
Read-only Jetson health snapshot for identity, memory, GPU, thermal, power, storage, services, and top processes.
mochi-remind
Handle due reminders — notify the user with natural language and mark them done.
dogfood
Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.