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 bks-lab/open-bridge --skill cmuxgit clone --depth 1 https://github.com/bks-lab/open-bridgeWrote 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/bks-lab/open-bridge/cmux)<a href="https://agentmods.dev/skills/bks-lab/open-bridge/cmux"><img src="https://agentmods.dev/badge/skills/bks-lab/open-bridge/cmux/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/bks-lab/open-bridge/cmux"><img src="https://agentmods.dev/badge/skills/bks-lab/open-bridge/cmux.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 | $0.00065 | $0.02263 |
| Opus 5.5 | $0.00026 | $0.00905 |
| Sonnet 5.5 | $0.00013 | $0.00453 |
| Haiku 4.5 | $0.00006 | $0.00226 |
Grade A, and why
cmux 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 6d 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 — 167 lines — stays where its author put it; the contents beside it link to each section on GitHub.
cmux
Manage cmux workspaces, tabs and the agent sessions inside them from one terminal, and bring them back when cmux loses them.
Optional. Everything here needs the cmux CLI, found as $CMUX_BIN, else
cmux on PATH, else the CLI inside the macOS app bundle (not on PATH under a
service manager). Without it the skill does
nothing: the scripts exit with a clear message and the workplace driver answers an
error, which scripts/workplace.py reads as "tab view unknown", never as "no tabs
open". Nothing else in the Bridge depends on cmux.
Decision tree
User wants to...
├── open a tab / split / workspace, spawn or watch an agent, sidebar, wait-for
│ → vocabulary table below, then references/local.md
├── "workspaces lost", "which sessions were running", cmux crashed or relaunched
│ → references/sessions.md: cmux_layout.py list → show → restore
├── one tab per task, "status of all tabs", "send <tab>: ..."
│ → scripts/workplace.py (CORE plan, docs/workplace.md) with this skill as driver
├── which session runs where, or what an agent is doing right now
│ → references/sessions.md: cmux sessions --json, vault search, agent journal
├── restart the agent in every tab (after an update)
│ → references/sessions.md § Restart
└── watch or answer agents on another machine, from a laptop or phone
→ references/remote.md
The workplace driver
scripts/workplace.py plans the day as workspaces with one agent tab per task
and knows no terminal. This skill ships the cmux side of it. Wire it in
bridge-config.yaml:
workplace:
driver: {command: ["python3", "${root}/skills/cmux/scripts/cmux_driver.py"]}
Then python3 scripts/workplace.py open --yes [--here "$CMUX_SURFACE_ID"] opens
the plan in cmux, and status, send, adopt steer the tabs.
python3 scripts/workplace.py launch --items A,B --yes [--here "$CMUX_SURFACE_ID"] [--target tab|workspace|area|auto] opens one tab per chosen task or inbox item
(docs/workplace.md § Launching chosen items). What the driver guarantees:
What ships with it
10 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.
- references/local.md 17 KB
- references/remote.md 3.3 KB
- references/sessions.md 7.6 KB
- scripts/cmux_driver.py 34 KB runs code
- scripts/cmux_layout.py 29 KB runs code
- scripts/cmux-claude-restart.sh 14 KB runs code
- scripts/cmux-open.sh 5.1 KB runs code
- scripts/spawn-workspace.sh 4.3 KB runs code
- tests/test_cmux_driver.py 40 KB runs code
- tests/test_cmux_layout.py 22 KB runs code
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.
- 6d ago Changed · +12 lines ac076e8e56b5
- 11d ago First seen · 155 lines · 65 tokens per session scan A f22114255d08
cmux is a skill published in the GitHub repository bks-lab/open-bridge (13 stars, last pushed today), licensed MIT. It adds 65 tokens to every session and 2,263 once invoked, about $0.0003 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 2026-10-05.
Other skills, from other repositories
slack
Use the assigned Slack bot from Slack conversations, Paperclip tasks, and routines to read shared discussions and collaborate.
agentmail
Use your assigned AgentMail inbox to read email tasks, explicitly send or reply, and check delivery. Provided automatically by your inbox assignment.
kungfu-agent-onboarding
Use when a user asks to understand, start, inspect, extend, or safely operate installed Kungfu; verify the installed pack, select one intent route, personalize the explanation, and propose one smallest safe next action.
cog-braindump-capture
Capture raw thoughts with automatic domain classification and vault routing.
context-preservation
State capture and restore across context window compactions. Monitors usage thresholds and serializes quality, task, and spec state for seamless continuation.
cog-knowledge-consolidation
Build structured knowledge frameworks from scattered vault notes with source attribution.