A procedure for spawning and coordinating worker AI sessions through the crewd system. It covers starting a worker, checking its progress, and collecting its result.

Use it when splitting work among AI agents, choosing an agent engine or profile, monitoring worker sessions, or passing messages and results between agents.

Background workers may finish without sending a completion notice, and repeated requests can create duplicate work. The procedure provides a defined way to track each worker and retrieve its output.

crew — Skill for Claude Code and Codex by DioNanos

Install

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.

Any agent

npx skills add DioNanos/mcp-crewd-rs --skill crew

Clone the repo

git clone --depth 1 https://github.com/DioNanos/mcp-crewd-rs

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/dionanos/mcp-crewd-rs/crew/github.svg)](https://agentmods.dev/skills/dionanos/mcp-crewd-rs/crew)

Your own site

<a href="https://agentmods.dev/skills/dionanos/mcp-crewd-rs/crew"><img src="https://agentmods.dev/badge/skills/dionanos/mcp-crewd-rs/crew/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/dionanos/mcp-crewd-rs/crew"><img src="https://agentmods.dev/badge/skills/dionanos/mcp-crewd-rs/crew.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 90 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,263 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00090	$0.01263
Opus 5	$0.00045	$0.00632
Sonnet 5	$0.00018	$0.00253
Haiku 4.5	$0.00009	$0.00126

Measured 11d ago against content hash cba59380c1b1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

crew 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 11d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (scripts/crew-mcp-driver.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

skills/crew/SKILL.md · 96 lines

How it starts

The opening of the file, as written. The whole thing — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.

crew — driving worker cells over the crewd fabric

crewd is a cell-fabric daemon: one AI session (the coordinator) spawns and coordinates other AI worker sessions (cells) through a single MCP server over a local Unix socket. This skill is the operational layer the tool schemas do not carry — the procedure and the gotchas.

Mental model

A cell is a worker AI session. Its identity is its working directory: the spawned engine loads whatever CLAUDE.md, .mcp.json, memory and skills live in the cwd you pass. One daemon serves many personas by spawning in different project roots.
Engines: claude (Agent SDK shim), codex (codex app-server), pi.
Profiles (claude): max (host credentials) or any profile declared in crewd.toml [profile.<name>]. Passed as a free string to cell_spawn.
Spawning is fire-and-forget: cell_spawn returns a crewd_thread_id immediately and the turn runs in the background.

The core loop: spawn → poll → result

Spawn

cell_spawn { engine, profile?, cwd, task, idempotency_key, mode:"background" } → returns { crewd_thread_id, replayed } at once. Always pass a unique idempotency_key (re-spawning with the same key returns the same thread, replayed:true).

Poll — there is NO completion push. The coordinator MUST poll; a finished worker notifies nobody. Call cell_status { crewd_thread_id } on an interval and watch state:

state	meaning
`spawning`	accepted, engine starting
`running`	turn in progress
`idle`	turn completed (engine went down) — this is the normal "done"
`timeout`	the turn exceeded the per-turn timeout (may be partial)
`failed` / `interrupted` / `failed_unknown`	turn/engine failure

Gotcha: poll for the state leaving running/spawning. Do not wait for a state literally named finished/succeeded — a normal completion surfaces as idle, and a slow worker as timeout. A poller that only matches finished never fires.

Read the full file on GitHub · 96 lines

Files

What ships with it

1 file 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.

scripts/crew-mcp-driver.py 4.5 KB runs code

Changes

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.

11d ago First seen · 96 lines · 90 tokens per session scan A cba59380c1b1

Subscribe to this mod's changes

crew is a skill published in the GitHub repository DioNanos/mcp-crewd-rs (0 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 90 tokens to every session and 1,263 once invoked, about $0.0005 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 2026-08-31.

Building agents agent ai-agents claude codex daemon mcp

b00t-interface-library

Design and implement a Rust interface library in l3dg3rr that acts as a feature-configurable lifecycle manager for b00t processes. The library compliantly implements init → operate → terminate → lifecycle maintenance of miscellaneous process surfaces (MCP servers, daemons, sidecars) with deterministic governance…

elasticdotventures/_b00t_ · 94 tokens

crew

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

crew — driving worker cells over the crewd fabric

Mental model

The core loop: spawn → poll → result

What ships with it

What this file has done since we first saw it

Other skills, from other repositories

b00t-interface-library

ensembl-database

mochi-remind

aatmf-t10-confidentiality-breach

memory

omh-code-review