Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/manaflow-ai/cmux-skillsnpx agentmods add skills/manaflow-ai/cmux-skills/cmux-cliWrote 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/manaflow-ai/cmux-skills/cmux-cli)<a href="https://agentmods.dev/skills/manaflow-ai/cmux-skills/cmux-cli"><img src="https://agentmods.dev/badge/skills/manaflow-ai/cmux-skills/cmux-cli/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/manaflow-ai/cmux-skills/cmux-cli"><img src="https://agentmods.dev/badge/skills/manaflow-ai/cmux-skills/cmux-cli.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Socket pass
- Snyk pass
- NVIDIA SkillSpector warn
SkillSpector: 4 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 90 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 250 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Rogue Agent · line 137 Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
- medium Agent Snooping · line 254 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
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 | $0.00057 | $0.02447 |
| Opus 5 | $0.00028 | $0.01223 |
| Sonnet 5 | $0.00011 | $0.00489 |
| Haiku 4.5 | $0.00006 | $0.00245 |
Grade A, and why
cmux-cli 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 10d 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 — 259 lines — stays where its author put it; the contents beside it link to each section on GitHub.
cmux CLI
Use this skill when a task is best handled through the cmux command line, or when the user asks how to use, inspect, script, or document the cmux CLI. Prefer the live CLI help for exact syntax, then apply the safety rules here.
Prerequisites
Use the cmux binary on PATH for normal user workflows:
cmux --help
cmux version
cmux ping
When dogfooding a tagged Debug build from a cmux source checkout, use the tag-bound helper instead of /tmp/cmux-cli:
CMUX_TAG=<tag> scripts/cmux-debug-cli.sh --help
CMUX_TAG=<tag> scripts/cmux-debug-cli.sh identify --json
scripts/cmux-debug-cli.sh targets /tmp/cmux-debug-<tag>.sock, uses the matching CLI from the tagged app bundle, and scrubs ambient cmux terminal context before running.
Discovery
Always discover current syntax from the CLI before giving exact command help:
cmux --help
cmux help
cmux <command> --help
cmux docs settings
cmux docs shortcuts
cmux docs api
cmux docs browser
cmux docs agents
cmux docs dock
Some subcommands print the top-level help instead of detailed subcommand help. If that happens, search the source locally:
rg -n "Usage: cmux <command>|case \"<command>\"|run.*<Command>" CLI
Do not fetch source files with gh api. Read the local checkout or active worktree.
Mental Model
cmux exposes app state over a Unix socket.
- Window: top-level macOS cmux window.
- Workspace: sidebar tab-like container inside a window.
- Pane: split region inside a workspace.
- Surface: tab inside a pane. A surface can host a terminal, browser, markdown viewer, diff viewer, or other panel.
- Panel: lower-level content implementation. Prefer surface commands unless a command explicitly requires
--panel.
Handle inputs usually accept UUIDs, refs such as window:1, workspace:2, pane:3, surface:4, or numeric indexes. Output defaults to refs:
cmux --id-format refs identify
cmux --id-format both list-pane-surfaces --workspace workspace:1
cmux --json --id-format both tree --all
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.
- 10d ago First seen · 259 lines · 57 tokens per session scan A a8245cc2c7ca
cmux-cli is a skill published in the GitHub repository manaflow-ai/cmux-skills (31 stars, last pushed 2d ago), licensed MIT. It adds 57 tokens to every session and 2,447 once invoked, about $0.0003 per session on Opus 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
slack
Interact with Slack workspaces using browser automation. Use when the user needs to check unread channels, navigate Slack, send messages, extract data, find information, search conversations, or automate any Slack task. Triggers include "check my Slack", "what channels have unreads", "send a message to", "search Slack…
browser-use
Direct browser control via CDP for web interaction: automation, scraping, testing, screenshots, and site/app work.
Interceptor
Real Chrome/Brave + macOS Computer Use from inside the browser — zero CDP fingerprint, real sessions; mandatory for visual deploy verification. Drive clickwork yourself. USE WHEN verify deploy, confirm UI, screenshot verification, computer use, macos automation, debug web, troubleshoot, visual check, motion/animation…
cli-anything-macrocli
Use when the agent wants to define, list, inspect, or execute GUI macros via the MacroCLI CLI. Macros are parameterized, CLI-callable workflows — the agent invokes macro run and the system handles backend routing (plugin, file transform, accessibility, compiled GUI replay).
Apify
Scrapes social platforms, business data, and e-commerce via Apify actors — Instagram, LinkedIn, TikTok, YouTube, Facebook, Google Maps, Amazon, and web crawls — filtering in code. USE WHEN scrape Instagram, scrape LinkedIn, scrape TikTok, scrape YouTube, scrape Facebook, Google Maps leads, Amazon reviews, business…
BrightData
4-tier progressive web scraping that auto-escalates WebFetch to curl to Interceptor to Bright Data proxy for bot detection and CAPTCHAs, with single-URL and multi-page crawl modes, output as markdown. USE WHEN Bright Data, scrape URL, web scraping, bot detection, crawl site, CAPTCHA, can't access, site blocking…