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 coldmint9/Synax --skill computer-usegit clone --depth 1 https://github.com/coldmint9/SynaxWrote 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/coldmint9/synax/computer-use)<a href="https://agentmods.dev/skills/coldmint9/synax/computer-use"><img src="https://agentmods.dev/badge/skills/coldmint9/synax/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.
<a href="https://agentmods.dev/skills/coldmint9/synax/computer-use"><img src="https://agentmods.dev/badge/skills/coldmint9/synax/computer-use.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00021 | $0.00366 |
| Opus 5.5 | $0.00008 | $0.00146 |
| Sonnet 5.5 | $0.00004 | $0.00073 |
| Haiku 4.5 | $0.00002 | $0.00037 |
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 21d 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
Computer Use
Cua controls the user's actual desktop, including signed-in applications and sensitive information. The built-in browser.* tools instead manage Synax's isolated Chromium. Use browser.* for ordinary web application testing, and Cua only when the user asks for their actual desktop, native apps, system dialogs, or signed-in browser.
- Observe first: list apps and windows, then get a fresh window state or accessibility tree. Never guess a window identifier or coordinate.
- Prefer fresh semantic elements and snapshot-bound tokens. Use pixels only when the semantic path cannot address the target. Do not invent
capture_idorelement_tokenvalues. - On macOS 27, use
delivery_mode: "foreground"for native window actions; background delivery is not reliable for the current CUA runtime. On other platforms, use background only when the runtime explicitly supports it. - Perform one consequential action at a time. Reobserve after window changes, page navigation, and actions; verify the user's desired postcondition before claiming success.
- A timeout or lost response may mean the action already happened. Never repeat click, typing, submit, launch, kill, or clipboard write without reobserving and user approval when appropriate.
- Avoid unnecessary full-screen captures, clipboard reads, secrets, and account data. Do not copy desktop content into logs or unrelated tools.
- If the driver is unavailable, permissions are missing, or a target is ambiguous, stop and explain the specific blocker rather than using shell commands to bypass restrictions.
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.
- 21d ago Changed b62fc07993de
- 22d ago First seen · 21 lines · 21 tokens per session scan A 680ecdeb5ea4
computer-use is a skill published in the GitHub repository coldmint9/Synax (13 stars, last pushed yesterday), licensed Apache-2.0. It adds 21 tokens to every session and 366 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 2026-09-29.
Other skills, from other repositories
computer-use
Read and drive native desktop apps through the accessibility layer — list apps, snapshot a window as a numbered element tree, then click/type/scroll/drag by element index or screen coordinates. Use for a desktop app, not a web page. macOS + Windows; off unless enabled in Settings.
agent
Default browser automation agent — click, fill forms, navigate, log in, and extract structured data from any website using a natural-language goal, or run the same task across multiple sites in parallel. New users get 600 free automation credits to start; beyond that it draws on your plan's automation credits (unlike…
computer-use
Choose Memmy browser, desktop Computer Use, or a structured connector for a live task on macOS or Windows 11; keep Computer History retrieval separate from UI actions.
telegram-userbot
Use when full MTProto control of Telegram account via Telethon. DM, Voice Note, Call, Video Call, Group/Channel management, member scraping, bot cloning, outreach automation, broadcast, CRM tracking, content reposting, scheduled messaging, webhook triggers. Use for all Telegram automation as a real user (not bot API).
recovery
Structured error recovery for failed or ineffective browser.
axstream
Fast deterministic macOS UI automation via replayable macros. Use BEFORE driving any macOS app's UI by hand (clicking/typing via computer-use or screenshots) — if a macro exists, replay finishes the task in seconds with no per-step reasoning; after doing a UI task manually, save it as a macro so the next run is fast.…