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 agentmods add instructions/craigsc/cmux/claude-mdgit clone --depth 1 https://github.com/craigsc/cmuxWrote 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/instructions/craigsc/cmux/claude-md)<a href="https://agentmods.dev/instructions/craigsc/cmux/claude-md"><img src="https://agentmods.dev/badge/instructions/craigsc/cmux/claude-md.svg" alt="Measured on agentmods" 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 | $0.00673 | $0.00673 |
| Opus 5 | $0.00336 | $0.00336 |
| Sonnet 5 | $0.00135 | $0.00135 |
| Haiku 4.5 | $0.00067 | $0.00067 |
Grade A, and why
cmux CLAUDE.md scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- `install.sh` — curl-pipe installer that downloads cmux.sh to `~/.cmux/` and adds source line to shell RC How it starts
The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What is cmux
cmux is a pure Bash shell tool that manages git worktree lifecycles for running parallel Claude Code sessions. Each worktree gets its own isolated working directory so multiple Claude agents can work on the same repo simultaneously.
Architecture
Single-file shell tool — all logic lives in cmux.sh (~560 lines), which is sourced into the user's shell (bash/zsh). There is no build step, no dependencies, and no test suite.
Function structure:
cmux()— public dispatcher, routes subcommands to_cmux_<cmd>functions_cmux_*()— private functions:_cmux_new,_cmux_start,_cmux_cd,_cmux_ls,_cmux_merge,_cmux_rm,_cmux_rm_all,_cmux_init- Helper functions:
_cmux_repo_root,_cmux_safe_name,_cmux_worktree_dir,_cmux_spinner_start/stop
Key design patterns:
- Idempotent operations (e.g.
cmux newreuses existing worktrees) - Context-aware:
cmux rmandcmux mergewith no args detect the current worktree from$PWD - Branch name sanitization:
feature/foo→feature-foofor directory names - Worktrees live under
.worktrees/<safe-branch-name>/in the repo root
Setup hook system: .cmux/setup is an executable bash script that runs after worktree creation. It handles project-specific init (symlink secrets, install deps, codegen). cmux init uses Claude CLI (claude -p) to auto-generate this hook.
Key files
cmux.sh— the entire applicationinstall.sh— curl-pipe installer that downloads cmux.sh to~/.cmux/and adds source line to shell RC.cmux/setup— project-specific worktree setup hook (committed to repo)examples/setup-node— example setup hook for Node.js projects
Shell conventions
- 2-space indentation
- Functions prefixed
_cmux_for internal,cmuxfor public - Zsh compatibility: uses
setopt localoptions nomonitorwhere needed for job control - Error handling: validate inputs, check
command -v, guard with|| return 1 - No
set -ein cmux.sh (it's sourced, not executed); install.sh usesset -e
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.
- 5d ago First seen · 50 lines · 673 tokens per session scan A 796eb0e0f181
cmux CLAUDE.md is an instructions file published in the GitHub repository craigsc/cmux (603 stars, last pushed 2mo ago), licensed MIT. It adds 673 tokens to every session, about $0.0034 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
agent-of-empires AGENTS.md
AGENTS.md instructions for agent-of-empires/agent-of-empires, covering repository guidelines, where to look, commands, code rules and tests.
terminal-browser AGENTS.md
AGENTS.md instructions for zenbu-labs/terminal-browser, a project described as: A browser inside your terminal.
obsidian-claude-sidebar CLAUDE.md
Claude Code instructions for derek-larson14/obsidian-claude-sidebar, covering claude sidebar - agent guide, what this plugin does, two things you can help the user do, step 1: diagnose and step 2a: file a bug report.
mandible AGENTS.md
AGENTS.md instructions for AS-FOSS/mandible, covering agents.md — working agreements for ai agents in this repo, 1. the invariant that defines the project, 2. hard limits, the scripts and 3. rules no lint can catch.
glances CLAUDE.md
Claude Code instructions for nicolargo/glances, covering claude.md — glances maintainer context, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.
liney AGENTS.md
AGENTS.md instructions for everettjf/liney, covering liney repository collaboration guide, 交流语言 / communication language, project overview, repository layout and build and test.