Borrowing it
Nothing to install: this file belongs to disler/learning-cmux-with-agents. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/disler/learning-cmux-with-agents/main/.claude/commands/cmux-did-spawn.mdgit clone --depth 1 https://github.com/disler/learning-cmux-with-agentsWrote 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/commands/disler/learning-cmux-with-agents/cmux-did-spawn)<a href="https://agentmods.dev/commands/disler/learning-cmux-with-agents/cmux-did-spawn"><img src="https://agentmods.dev/badge/commands/disler/learning-cmux-with-agents/cmux-did-spawn/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/commands/disler/learning-cmux-with-agents/cmux-did-spawn"><img src="https://agentmods.dev/badge/commands/disler/learning-cmux-with-agents/cmux-did-spawn.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 | $0.00052 | $0.00947 |
| Opus 5 | $0.00026 | $0.00474 |
| Sonnet 5 | $0.00010 | $0.00189 |
| Haiku 4.5 | $0.00005 | $0.00095 |
Grade A, and why
cmux-did-spawn 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 9d 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 — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
cmux Did Spawn
Purpose
A full-stack Flotion team was already booted as a workspace in a cmux window
by the fastcc/fastpi recipe (declarative layout — all 5 pi agents launched at
once). That window may be shared with other teams — each team is its own
workspace, so locate this team's workspace by name, not by position. Your job is
to take command of it: read the spawn file, find the window and the team's
workspace, confirm the agents are up, and stand ready to drive the lead. You
are the orchestrator and you run in a terminal, outside cmux.
Variables
SPAWN_FILE: $1 # path to the .team/.spawn.json written at spawn time
Instructions
- cmux short refs are positional and renumber. Never cache a
surface:Nacross time. The only stable handle is the window UUID in the spawn file — locate the team by it and rediscover surface refs at the moment you use them, scoped to the team's workspace. - Talk to the LEAD only. The lead dispatches its own workers (plan, build-be, build-fe, test). You hand the lead a feature; it does the rest.
- The team is already booting — don't re-spawn anything. If a pane shows a shell instead of an agent, that one pi failed to start; report it, don't recreate the whole team.
- Confirm cmux is reachable first (
cmux identify --json).
Workflow
- Read
SPAWN_FILE:F=$(jq -r .feature "$SPAWN_FILE"); WIN=$(jq -r .window "$SPAWN_FILE") WSNAME=$(jq -r '.workspace_name // .feature' "$SPAWN_FILE") - Confirm cmux is reachable (
cmux identify --json); if not, tell the user toopen -a cmuxand stop. - Locate this team's workspace inside the (possibly shared) window by its
name, not by position — other teams may be sibling workspaces in the same
window. Then map each role to its current surface ref by the layout name:
Capture the surface ref on the line namedWS=$(cmux workspace list --window "$WIN" --json \ | jq -r --arg n "$WSNAME" '.workspaces[] | select(.custom_title==$n) | .ref' | head -1) cmux list-pane-surfaces --workspace "$WS" # names: lead / plan / build-be / build-fe / testlead— that's your dispatch target. - Confirm the team is alive: read the lead pane once
(
cmux read-screen --surface <lead> --scrollback --lines 30). The workers should each be replyingready: <role>. - You are now oriented. Wait for the user to give you a feature request (they may also give it directly to the leads), then:
cmux send --surface <lead> "<feature>"; cmux send-key --surface <lead> enter cmux read-screen --surface <lead> --scrollback --lines 60 # watch it coordinate - Now follow the
Reportsection.
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.
- 9d ago First seen · 80 lines · 52 tokens per session scan A fbb307b53e36
cmux-did-spawn is a command published in the GitHub repository disler/learning-cmux-with-agents (109 stars, last pushed 2mo ago), licensed MIT. It adds 52 tokens to every session and 947 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 commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.