Agent S is an open framework for AI agents that interact with graphical computer interfaces through actions such as a person would take. It is for building computer-use agents that plan tasks, retrieve relevant information, and operate desktop or mobile environments.
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 skills add simular-ai/Agent-S --skill openclawgit clone --depth 1 https://github.com/simular-ai/Agent-SWrote 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/simular-ai/agent-s/openclaw)<a href="https://agentmods.dev/skills/simular-ai/agent-s/openclaw"><img src="https://agentmods.dev/badge/skills/simular-ai/agent-s/openclaw/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/simular-ai/agent-s/openclaw"><img src="https://agentmods.dev/badge/skills/simular-ai/agent-s/openclaw.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00000 | $0.00760 |
| Opus 5.5 | $0.00000 | $0.00304 |
| Sonnet 5.5 | $0.00000 | $0.00152 |
| Haiku 4.5 | $0.00000 | $0.00076 |
Grade A, and why
openclaw 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 — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent-S - Autonomous GUI Agent
Agent-S is a powerful autonomous agent that can control your computer's graphical interface to complete complex tasks. It combines vision and action understanding to interact with any GUI element.
What It Does
Agent-S can:
- Navigate and interact with desktop applications
- Fill forms, click buttons, and manipulate GUI elements
- Complete multi-step workflows across different applications
- Take screenshots and understand visual interfaces
- Execute complex GUI automation tasks autonomously
When to Use
Use Agent-S when you need to:
- Automate GUI-based tasks that don't have CLI alternatives
- Interact with desktop applications programmatically
- Complete workflows that require visual understanding
- Perform actions across multiple applications
- Test GUI interfaces
How to Invoke
Call the Agent-S wrapper via bash from the OpenClaw skills directory:
./agent_s_task "task description"
Or if installed in the default OpenClaw skills location:
~/.openclaw/workspace/skills/agent-s/agent_s_task "task description"
Note: Agent-S tasks can take 2-5 minutes to complete (up to 15 steps by default). The wrapper will wait for completion.
Parameters
task(required): Natural language description of the GUI task to completemax_steps(optional): Maximum steps the agent can take (default: 15)enable_reflection(optional): Enable self-reflection for better performance (default: true)
Examples
# Basic navigation
agent_s_task(task="Open Finder and create a new folder called 'Reports'")
# Form filling
agent_s_task(task="Open TextEdit, create a new document, and type 'Hello World'")
# Multi-step workflows
agent_s_task(task="Open Chrome, search for 'Python tutorials', and bookmark the first result")
# Application interaction
agent_s_task(task="Open System Preferences and check the current display resolution")
Technical Details
Agent-S uses:
- Main Model: Claude Sonnet 4.5 for reasoning and planning
- Grounding Model: UI-TARS-1.5-7B for visual grounding and coordinate extraction
- Screen Resolution: Automatically scaled to 2400px max dimension
- Platform Support: macOS, Linux, Windows
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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 First seen · 103 lines · 0 tokens per session scan A e42a63f84807
openclaw is a skill published in the GitHub repository simular-ai/Agent-S (12,629 stars, last pushed 8d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 760 tokens. 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
browserwing-executor
Control browser automation through HTTP API. Supports page navigation, element interaction (click, type, select), data extraction, accessibility snapshot analysis, screenshot, JavaScript execution, and batch operations.
pixelbrowse
Screenshot and visually read any web page or document using pixelshot. Use instead of fetching raw HTML when you need to see what a page looks like, read visual content (charts, diagrams, infographics), check layouts, or verify UI. Triggers: "look at this page", "screenshot", "what does this site look like", "check…
dev-browser
Browser automation with persistent page state. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include "go to [url]", "click on", "fill out the form", "take a screenshot", "scrape", "automate", "test the website"…
memory-bridge
Browse or compare wiki knowledge by originating AI tool. Use for cross-tool memory questions, tool-specific knowledge, or gap/diff analysis between Codex, Claude, Hermes, and other agents.
serper-search
Search the public web through Serper instead of the built-in WebSearch tool. Use when an agent needs current information, general web sources, or URLs not available from local files and project APIs. SERPERAPIKEY is available in the environment.
browserwing-admin
Manage and operate BrowserWing — an intelligent browser automation platform. Install dependencies, configure LLM, create/manage/execute automation scripts, use AI-driven exploration to generate scripts, browse the script marketplace, and troubleshoot issues.