--- name: claude-computer-use-1968ec12cad0 description: >- Claude computer use is an Anthropic API tool that lets Claude see a desktop through screenshots and operate it with mouse and keyboard. Use when automating GUI workflows, browser automation without CSS selectors, desktop app testing, or any task that requires seeing and interacting with a graphical interface. license: Apache-2.0 compatibility: "Anthropic API key, Python 3.9+, a sandboxed Linux desktop (Xvfb) or VM" metadata: author: terminal-skills version: "1.1.0" repository: https://github.com/anthropics/anthropic-quickstarts category: data-ai tags: "computer-use, claude, anthropic, automation, gui" use-cases: "Automate form filling in legacy web applications without APIs, Run end-to-end GUI tests on desktop applications, Extract data from applications that have no API or export feature" agents: "claude-code, cursor" --- # Claude Computer Use ## Overview Computer use is a client-side tool of the Claude API. Claude asks for actions (take a screenshot, click, type, scroll), your program performs them on a desktop you control and returns the result, and the loop repeats until the task is done. No selectors are needed, so it suits legacy apps and GUIs with no API. For tasks that stay inside web pages, Anthropic's browser use tool is the closer fit. Current API shape (checked 2026-10): a single entry `{"type": "computer_toolset_20260801"}` in `tools`, no beta header. Claude 5.5 and later models on the Claude API and Google Cloud accept only this toolset. Older models (Opus 4.5 to 4.8, Sonnet 4.6) use `computer_20251124` with beta header `computer-use-2025-11-24`; Opus 5 and Sonnet 5 accept both shapes. Sonnet 4.5 and Haiku 4.5 use `computer_20250124` with `computer-use-2025-01-24`. The `computer_20241022` version from the first beta is gone. **Always run it in a sandbox (container or VM) without credentials or production access.** Web pages and images can contain prompt injections. ## Instructions ### Setup ```bash pip install anthropic pillow pyautogui sudo apt-get install -y xvfb scrot xdotool # Linux sandbox display and screenshots export ANTHROPIC_API_KEY=... # from the environment, never in code ``` The fastest safe start is Anthropic's reference container (Linux desktop, VNC viewer, Streamlit UI, agent loop): ```bash docker run -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \ -v $HOME/.anthropic:/home/computeruse/.anthropic \ -p 127.0.0.1:5900:5900 -p 127.0.0.1:8501:8501 -p 127.0.0.1:6080:6080 -p 127.0.0.1:8080:8080 \ -it ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest ``` Open http://localhost:8080 for the combined view. Source: github.com/anthropics/anthropic-quickstarts, `computer-use-demo`. ### Declare the tools ```python import anthropic client = anthropic.Anthropic() MODEL = "claude-sonnet-5-5" TOOLS = [ {"type": "computer_toolset_20260801"}, {"type": "bash_20250124", "name": "bash"}, {"type": "text_editor_20250728", "name": "str_replace_based_edit_tool"}, ] ``` The toolset takes no display size: coordinates are in the pixel space of the screenshots you return. It rejects `name`, `display_width_px`, `display_height_px`, `display_number` and `enable_zoom`. To withhold a member, use `"configs": {"zoom": {"enabled": false}}`. Declaring the toolset costs about 4,500 input tokens per request. ### Member tools Claude's `tool_use` blocks have `"toolset_name": "computer"` and a `name` that is one of 17 members; `input` has no `action` field. | Member | Input | |---|---| | `screenshot`, `cursor_position` | none | | `zoom` | `region: [x0, y0, x1, y1]` (returns an image) | | `left_click`, `right_click`, `middle_click`, `double_click`, `triple_click` | optional `coordinate: [x, y]`, optional `text` modifiers such as `shift` or `ctrl+shift` | | `left_click_drag` | `start_coordinate`, `coordinate` | | `mouse_move` | `coordinate` | | `left_mouse_down`, `left_mouse_up` | none | | `scroll` | `scroll_direction` (up/down/left/right), `scroll_amount`, optional `coordinate` | | `type` | `text` | | `key` | `text` such as `ctrl+s`, optional `repeat` 1-100 | | `hold_key` | `text`, `duration` seconds (max 300) | | `wait` | `duration` seconds (max 300) | ### Agent loop ```python import base64, subprocess, time from pathlib import Path import pyautogui def screenshot_block() -> dict: subprocess.run(["scrot", "-o", "/tmp/screen.png"], check=True) # Images must fit the model limit (2576 px long edge for 5.x models): downscale # larger screens with Pillow and scale Claude's coordinates back up. data = base64.standard_b64encode(Path("/tmp/screen.png").read_bytes()).decode() return {"type": "image", "source": {"type": "base64", "media_type": "image/png", "data": data}} def run_member(name: str, a: dict): """Perform one computer member call; return text or an image block.""" xy = a.get("coordinate") if name == "screenshot": return [screenshot_block()] if name == "left_click": pyautogui.click(*xy) elif name == "right_click": pyautogui.rightClick(*xy) elif name == "double_click": pyautogui.doubleClick(*xy) elif name == "mouse_move": pyautogui.moveTo(*xy) elif name == "type": pyautogui.write(a["text"], interval=0.02) elif name == "key": for _ in range(a.get("repeat", 1)): pyautogui.hotkey(*a["text"].lower().split("+")) elif name == "scroll": amount = a["scroll_amount"] * (-1 if a["scroll_direction"] == "down" else 1) pyautogui.scroll(amount, *(xy or pyautogui.position())) elif name == "wait": time.sleep(a["duration"]) else: raise NotImplementedError(name) # implement the rest as needed (zoom, drag, ...) time.sleep(0.5) return "OK" def run_agent(task: str, max_steps: int = 25) -> str: messages = [{"role": "user", "content": task}] for _ in range(max_steps): resp = client.messages.create(model=MODEL, max_tokens=4096, tools=TOOLS, messages=messages) messages.append({"role": "assistant", "content": resp.content}) calls = [b for b in resp.content if b.type == "tool_use"] if not calls: return next((b.text for b in resp.content if b.type == "text"), "") results, failed = [], False for call in calls: # a turn may hold a batch: run in order res = {"type": "tool_result", "tool_use_id": call.id} if getattr(call, "toolset_name", None) == "computer": res["toolset_name"] = "computer" if failed: res.update(is_error=True, content="Not executed: an earlier computer action in this turn failed.") else: try: if res.get("toolset_name") == "computer": out = run_member(call.name, call.input) else: out = run_other_tool(call) # your bash / editor handlers res["content"] = out except Exception as exc: failed = True res.update(is_error=True, content=f"Error: {exc}") results.append(res) messages.append({"role": "user", "content": results}) return "Stopped: max_steps reached" ``` Rules from the docs: answer every `tool_use` block (an unanswered one is a 400 error), echo `toolset_name` on computer results, stop at the first failure and answer the rest with the exact halt text above, and return an image only for `screenshot` and `zoom` (text `OK` otherwise). `run_other_tool` is yours to write (run the bash command in the sandbox, apply the editor command). With `tool_choice` set to `{"type": "auto", "disable_parallel_tool_use": true}` Claude issues one action per turn. ### Migrating from computer_20251124 or computer_20241022 Remove the beta header, change the tool entry to the toolset, dispatch on `name` plus `toolset_name` instead of `input.action`, iterate over every `tool_use` block, honor `repeat` on `key`, resize screenshots yourself, and decide whether to keep `zoom` (on by default, whereas `enable_zoom` defaulted to off). ### Human approval Ask a person before irreversible steps. Check before each block runs, because a batch can finish a multi-step action in one turn: ```python RISKY_WORDS = ("submit", "delete", "purchase", "send", "pay") def approved(call, task: str) -> bool: if call.name in ("left_click", "double_click", "key") and any(w in task.lower() for w in RISKY_WORDS): return input(f"Allow {call.name} {call.input}? [y/N] ").strip().lower() == "y" return True ``` ## Examples ### Fill in a legacy web form User: "Open the supplier portal in Firefox and register Brightwell Tools, contact maria.lopez@brightwelltools.com. Stop before pressing Submit." ```python print(run_agent( "Firefox is open on the supplier portal at http://localhost:8080/register. " "Fill Company='Brightwell Tools' and Email='maria.lopez@brightwelltools.com'. " "After each step take a screenshot and confirm it worked. Do not press Submit; " "tell me when the form is ready." )) ``` Result: a handful of `screenshot`, `left_click` and `type` calls, then Claude replies that the form is filled and waiting. Add the approval check before any Submit click. ### Read data from an app with no export User: "Copy the open invoice numbers from the desktop accounting app into a text file." ```python print(run_agent( "In the Ledger app, open Invoices > Open. Read every invoice number in the list " "(scroll if needed, zoom into the table to read small text) and write them one per line " "to /home/computeruse/open_invoices.txt using the bash tool. Done when the file exists." )) ``` Result: Claude zooms into the list, scrolls, writes the file with bash, and returns a summary with the count. ## Guidelines - Use a throwaway container or VM; never your daily desktop with signed-in browsers. Restrict network access to an allowlist of domains. - Put the task text before the screenshot in the first user turn; click accuracy improves. - Ask Claude to screenshot and verify after each step, and to use keyboard shortcuts for dropdowns and scrollbars that resist the mouse. - Always cap the loop (`max_steps`) and log actions and screenshots for audit. - Do not hand over passwords casually: logged-in sessions raise prompt-injection risk. Keep payment and admin accounts out of the sandbox. - Latency is high and coordinates can be wrong; it is not for speed-critical or precision-critical work. Prefer APIs, then Playwright, when they exist. - Each screenshot is billed as image input; keep screens near the model's size limit and trim history of old screenshots in long runs. - Everything runs on your side: screenshots and files stay in your environment; Anthropic only sees what you send in each request. --- Source attribution: terminal-skills Original project: https://github.com/TerminalSkills/skills/blob/8a9852d489f274538be9bf1c9121aeb403750b1c/skills/claude-computer-use/SKILL.md License: Apache-2.0 License information: https://www.apache.org/licenses/LICENSE-2.0 Original upstream instructions; downloaded from Mega AI Hub. ## Upstream license notice Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 1. Definitions. "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document. "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License. "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License. "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types. "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below). "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof. "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to the Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution." "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by the Licensor and subsequently incorporated within the Work. 2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. 3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed. 4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and (b) You must cause any modified files to carry prominent notices stating that You changed the files; and (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding any notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. 5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions. 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file. 7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License. 8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages. 9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability. END OF TERMS AND CONDITIONS Copyright 2025 Terminal Skills Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. # Mega AI Hub resource metadata Creator / publisher: terminal-skills Original project: https://github.com/TerminalSkills/skills/blob/8a9852d489f274538be9bf1c9121aeb403750b1c/skills/claude-computer-use/SKILL.md Attribution basis: Source repository record Fields of work: Customer Support, Development & Engineering Roles: Technical Support, QA / Test Engineer, DevOps / SRE Specialties: Testing, DevOps Tools: Anthropic, Pyautogui Country applicability: Country not specified Classification basis: Source category, role and specialization labels Instruction SHA-256: 238bcabc3292f3b1f27b289a8cb1886b5548836f79372a6d2496a93ce4e1dac7