computer-use

computer-use is a skill for Claude Code, Codex from nexus-research-lab/nexus. It costs 51 tokens per session (1,709 once invoked), scanned A, original, Apache-2.0.

A skill for operating native desktop applications on macOS or Windows through the host's computer-control interface. It is separate from browser automation and ordinary file or network operations.

In plain words
What is it for?
Use it when the user explicitly asks for help with a native desktop application, such as selecting an app window and performing an allowed action.
Why use it?
It provides a defined, permission-based way to inspect and act on desktop apps while requiring fresh observations before actions.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it when the user explicitly asks for help with a native desktop application, such as selecting an app window and performing an allowed action.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/nexus-research-lab/nexus/computer-use
Install

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.

Any agent
npx skills add nexus-research-lab/nexus --skill computer-use
Clone the repo
git clone --depth 1 https://github.com/nexus-research-lab/nexus

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for computer-use

README.md
[![agentmods](https://agentmods.dev/badge/skills/nexus-research-lab/nexus/computer-use/github.svg)](https://agentmods.dev/skills/nexus-research-lab/nexus/computer-use)
Your own site
<a href="https://agentmods.dev/skills/nexus-research-lab/nexus/computer-use"><img src="https://agentmods.dev/badge/skills/nexus-research-lab/nexus/computer-use/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.

agentmods 80×15 button for computer-use

Your own site · 80×15
<a href="https://agentmods.dev/skills/nexus-research-lab/nexus/computer-use"><img src="https://agentmods.dev/badge/skills/nexus-research-lab/nexus/computer-use.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 51 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,709 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00051 $0.01709
Opus 5.5 $0.00020 $0.00684
Sonnet 5.5 $0.00010 $0.00342
Haiku 4.5 $0.00005 $0.00171

Measured 2d ago against content hash d203210a0a17, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

computer-use 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.

skills/computer-use/SKILL.md · 57 lines

How it starts

The opening of the file, as written. The whole thing — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Computer Use

当用户要求你查看或操作当前电脑上的原生桌面应用时使用本 Skill。浏览器网页优先使用 Browser 能力;普通文件读写使用文件或文档能力。不要为了绕过能力边界调用 shell、文件系统、浏览器调试协议或外部 MCP。

入口

Computer Use 只通过当前回合的 nexus.command 使用。先读取固定能力与权限:

{"domain":"computer","action":"inspect"}

如果返回错误,说明能力未启用、宿主权限不足或 sidecar 不可用;把这个事实告诉用户,不要猜 endpoint、token、CLI 参数,也不要尝试重放动作。

标准流程

  1. 需要新的操作时先读取精确 contract:

    {"domain":"computer","action":"contract","operation":"<operation>"}
    
  2. 已知目标时使用 list_applications 并传 query(如 "Chrome" 或 "com.google.Chrome"),避免对整个应用目录分页。query 只筛选名称/ID,不选择目标或授予权限;再用 select_target 以返回的精确应用名称或稳定 application id 选择目标,明确请求所需 allowed_actions、allow_foreground_input 和 ttl_seconds,由宿主确认后生效。参数上限读取 get_computer.action_limits/action_schema。

    所有 invoke 都需要顶层 request_id,包括只读操作。例如 {"domain":"computer","action":"invoke","operation":"list_applications","request_id":"list-apps-1","input":{"query":"Chrome"}}。 operation 只放顶层,input 只包含该操作 contract 声明的参数,不再嵌套 operation。

  3. 首次选择多窗口应用时先调用 list_windows,再把返回的 window_ref 或精确标题传给 observe;单窗口应用可直接 observe。后续省略 window_selector 会沿用仍存在的已选窗口;原窗口消失且有多个新窗口时再选择。window_selector 是工具参数,不是要求用户打开窗口选择弹窗。

  4. 每次动作必须使用最新观察;首次调用 observe,后续可使用动作返回的新 observation。截图或无障碍树过期后重新观察。需要同应用菜单与弹窗时,在能力支持下使用 application_scene=true。 观察默认包含截图;点击、移动、拖拽和滚动必须使用带截图的观察,目标点均使用截图左上角为原点的 0..1000 归一化坐标;不要传图片像素。滚动必须给出目标 x、y,正向 delta_y 向下滚动。快捷键优先使用 mod(macOS Command,Windows Control),回车使用 enter。

  5. 动作使用 perform_action,把业务输入放在 contract 允许的 closed object 中;顶层生成新的 request_id。如果结果未知,复用同一个 request_id 重试以完成对账,不能换 ID 重放。

  6. 动作返回 observation 时直接读取它;未返回时重新 observe。observation_error 只代表后续读取失败,不能据此重做动作。缓存结果标记 fresh_observation_required 时重新观察。需要确认业务状态时调用 verify_state。 普通 observe 的 observation_id 和 elements 在结果顶层;perform_action 的新观察在 observation 内。截图是单独的图片内容,不要把它当作文本分页读取。

  7. 任务结束调用 close_target,丢弃本轮的 session、observation、element 和 pending mutation 状态。

安全边界

  • 只使用宿主返回的 opaque refs;不能提交 PID、窗口句柄、AX/UIA 对象、路径、socket、token 或任意 native handle。
  • 输入、文本和截图遵守 contract 返回的长度与数量上限。不要把敏感文本写入普通回复或日志。
  • 常规动作使用 select_target 已批准的会话范围;退出应用及宿主标记的敏感操作仍需单独审批。被拒绝或权限不足时停止并说明原因。审批后 stale 必须重新观察,不自动重放。
  • 只有明确返回 state=minimized 才使用 restore_window;window_status 来自观察前的窗口发现,unknown/offscreen、小尺寸或台前调度缩略图都不能证明最小化。台前调度恢复使用 focus_window。最小化窗口先用 include_screenshot=false, accessibility="disabled" 观察,再执行 restore_window,恢复后截图。
  • session_active=true 仅表示应用授权仍在,不保证已有可用于动作的 observation。观察失败或尚未观察时,先修复观察,不要反复 select_target。
  • stale_observation 不是应用授权到期。失败的动作尝试也可能消耗旧观察;重新观察后根据当前画面决定下一步,不能机械重放。只有 not_dispatched 才能确定该尝试未执行。若连续输入再次让新观察失效,等待用户空闲再继续,不循环 observe/action,也不要求重新授权。
  • macOS 后台窗口(包括台前调度缩略窗口)若返回 target_unavailable 且 recovery_action="focus_window",对同一 window_ref 先做一次 include_screenshot=false, accessibility="disabled" 的元数据观察,再在已批准的范围内执行一次 focus_window,然后重新截图和读取无障碍树。禁止用未成功的观察直接执行动作;一次恢复仍失败就报告具体错误,不能反复聚焦或重新授权。此流程不解除 user_active。
  • 应用未运行时先 list_launchable_applications 并传相同 query,再用返回的精确 launch_ref 调用 launch_application;启动许可不包含后续输入权限。若应用在发现后已被用户启动,按 discover_applications 指引重新查询运行应用并 select_target,不要重复 launch。
  • macOS 两次动作之间的人工输入会等待连续空闲 2 秒(单次最多等 5 秒),保留应用授权,但输入之前的观察失效。重新观察后可继续;busy 且 recovery_action="wait_for_idle_then_observe" 时最多重新观察并尝试一次,持续忙碌就说明正在等用户空闲,不能无限重试。
  • user_active 表示原生动作执行期间用户接管:停止自动动作并说明,不因等待或重新观察而恢复,不自动换 session。用户明确要求继续后再走新的 select_target 授权;可能已执行一部分的动作先对账。
  • session 过期重新 select_target;回合授权上限到达后开始新回合,不静默续期。
  • indeterminate 只用原 request_id 对账一次;仍未知就停止并报告,不通过新 ID、重开 target 或 launch_application 绕过待对账状态。
  • 每个 mutation 只对应一个稳定 request_id;目标、动作或输入改变时必须换 ID。
  • 不要自动点击系统权限弹窗。Screen Capture、Accessibility 或 Input Control 未授予时,说明用户需要在系统设置中授权后再试。

Read the full file on GitHub · 57 lines

Changes

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.

  1. 2d ago First seen · 57 lines · 51 tokens per session scan A d203210a0a17

Subscribe to this mod's changes

computer-use is a skill published in the GitHub repository nexus-research-lab/nexus (160 stars, last pushed 2d ago), licensed Apache-2.0. It adds 51 tokens to every session and 1,709 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 2026-11-05.

Related

computer-use compared with skills from other repositories

computer-use next to 6 skills from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
computer-usenexus-research-lab/nexusSkill1602d agoApache-2.0A51Claude Code, Codex
agent-initializationPrism-Shadow/penguin-harnessSkill3.2kyesterdayApache-2.0A79Claude Code, Codex
cronshibing624/agenticaSkill3602d agoApache-2.0A52Claude Code, Codex
ceo-setupsuyoumo/ClawProBenchSkill8292mo agoApache-2.0A60Claude Code, Codex
commitment-triagesuyoumo/ClawProBenchSkill8292mo agoApache-2.0A29Claude Code, Codex
routine-advisorsuyoumo/ClawProBenchSkill8292mo agoApache-2.0A19Claude Code, Codex
delegationsuyoumo/ClawProBenchSkill8292mo agoApache-2.0A25Claude Code, Codex