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 zavora-ai/computer-use-mcp --skill computer-use-recoverygit clone --depth 1 https://github.com/zavora-ai/computer-use-mcpWrote 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/zavora-ai/computer-use-mcp/computer-use-recovery)<a href="https://agentmods.dev/skills/zavora-ai/computer-use-mcp/computer-use-recovery"><img src="https://agentmods.dev/badge/skills/zavora-ai/computer-use-mcp/computer-use-recovery/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/zavora-ai/computer-use-mcp/computer-use-recovery"><img src="https://agentmods.dev/badge/skills/zavora-ai/computer-use-mcp/computer-use-recovery.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 | $0.00023 | $0.00462 |
| Opus 5.5 | $0.00009 | $0.00185 |
| Sonnet 5.5 | $0.00005 | $0.00092 |
| Haiku 4.5 | $0.00002 | $0.00046 |
Grade A, and why
computer-use-recovery 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 13d 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.
What it actually says
Focus / window recovery
When a mutating tool returns focus_failed or similar:
| suggestedRecovery | Action |
|---|---|
activate_window |
activate_window(requestedWindowId) then retry |
unhide_app |
unhide_app → wait → activate_window → retry |
open_application |
open_application → wait (wait_for_window for the app) → retry |
If a third-party app steals focus after activation, retry once with focus_strategy: "prepare_display", then unhide_app for any hiddenBundleIds when done (unhide_app is in every profile since 7.6.1; apps still hidden when the session ends are restored automatically).
Always re-resolve window_id via list_windows if the target window may have moved or closed. get_target shows the session target set by set_target and whether its window is still on screen; call set_target again after a window is replaced.
user_active
user_active is not a focus failure: the person was typing or moving the mouse within COMPUTER_USE_USER_IDLE_MS (default 4 s), and the call would have taken focus or posted input. Do not retry in a loop, and do not add force: true unless the user asked for the action.
- Retry with
delivery: "pid"(no activation, no cursor movement; keys usually reach background apps, clicks often do not), or waita few seconds and retry, or- Ask the user.
Capture helper errors (macOS)
helper_busy (another capture held the helper lock too long), helper_timeout (the helper was killed after the call timeout) and sck_timeout (ScreenCaptureKit stalled for 10 s) mean the capture failed, not that the screen is empty. Wait a moment and retry once; doctor reports agent_helper. Do not run two captures at once yourself: the server serialises them.
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.
- 13d ago Changed b4d40afd2bdb
- 16d ago Changed · +10 lines · +3 tokens per session b3aa9cdfef1f
- 1mo ago First seen · 19 lines · 20 tokens per session scan A 835b0a9622d0
computer-use-recovery is a skill published in the GitHub repository zavora-ai/computer-use-mcp (86 stars, last pushed today), licensed MIT. It adds 23 tokens to every session and 462 once invoked, about $0.0001 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
workflow-interaction
A guide for exploring and operating real web pages and desktop applications. It describes how to locate the correct window or page, inspect available controls, and confirm that an action succeeded.
oswright
Drive real desktop applications through the OSWright MCP server - click, type, read on-screen text, manage windows and fill forms - while keeping token cost low by tracking what changed on screen instead of sending a screenshot after every step. Use when the user asks to operate a desktop app, automate a GUI that has…
agent-desktop
Operate desktop applications on macOS, Windows and Linux through native accessibility trees, using the agent-desktop CLI. Use when an AI agent must see or act on a desktop GUI: click buttons, fill forms, read UI state, navigate menus, toggle checkboxes, scroll, drag, type text, press shortcuts, take screenshots…
Templates
Starter buttons for the composer and instructions for making your own.
desktop-gui-inspect
Full-desktop automation for targets that have no browser and no API at all — a legacy Java-based NMS client, a vendor's Windows-only configuration utility, a terminal emulator with no scriptable interface. Drives OpenClaw's ClawHub computer-use skill (Xvfb+XFCE virtual desktop, xdotool input automation) to read…
turix-mac
Computer Use Agent (CUA) for macOS automation using TuriX. Use when you need to perform visual tasks on the desktop, such as opening apps, clicking buttons, or navigating UIs that don't have a CLI or API.