Skip to content
Agentmart
Submit a tool

team-shape

Choose how many agents to spawn and who decides, starting from one agent.

Skill by oneiron-dev

IndexedApache-2.0
Write a review

Chalk Labs

Need help deploying team-shape?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/oneiron-dev-team-shape

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Choose how many agents to spawn and who decides, starting from one agent.

Use cases: agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/team-shape

What's inside

Skill name
team-shape
Description
Choose how many agents to spawn and who decides, starting from one agent.
License
Apache-2.0
Instructions
57 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Team shape

Models carry weak priors about collaboration, and a naive team often loses
to one agent. This skill teaches the planner (and any parent about to
spawn) how to organise a team. It is seeded from coordination research and
revised from receipts: measured payoff rows outweigh every rule below.

## The first lesson: solo first

Start from one agent with a clear brief. When one agent already succeeds
often on tasks of this type, adding agents tends to hurt: more chances to
drift, more context spent coordinating, slower convergence. A team earns
its width only when its measured payoff beats solo at equal budget — never
because the task "feels big".

## Spawning

Say what you need, separate from the prompt: `need` (required, one or two
lines), `care` (light, medium, high, xhigh, max), and `notes` for
constraints. Set `model`, `width`, `pattern` or `budget` only to steer: a
seed says where learning starts, a pin fixes the value. Everything unset
belongs to the planner.

## The patterns (who decides)

- Solo — one agent. Works well for: most tasks.
- CEO decides — the leader alone, after advice. The default team.
  Works well for: work with independent advice streams.
- Committee — a few deliberate, the chair decides. Works well for:
  researching what exists and refining. It narrows ideas, so it is late
  on generative work.
- Vote — members vote first, debate only on request. Works well for:
  judging between finished options.
- Market — a referee signed before the work decides (a test suite, a
  holdout, a pinned judge). Works well for: checkable work at width.
  No referee, no market.
- Rotation — roles rotate by turn. Works well for: reframing and open
  questions. Starts untested.

## Budget

The budget is an input to the plan, never a refusal. A 20-agent
tournament on a three-agent budget is pointless: fit the best shape the
budget allows and say what more would buy. Width has no count cap; the
spend cap is the cap. When a run nears the end of its lease, agents get a
wrap-up warning, partial results are saved in a resumable hold, and the
owner gets a notice. A suspension is a hold, not a kill.

## Care

Care spends on shape, not just effort. Light care rarely justifies a
team. High care and up is where a wider, stronger shape can pay — and
where the planner itself climbs its ladder: a cheap classifier for the
easy cases, a small thinking model next, a strong model for xhigh, max
or big plans.

From oneiron-dev/oneiron, Apache-2.0.

More from oneiron-dev/oneiron

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Oct 16, 2026 (26 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
Apache-2.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from oneiron-dev

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get team-shape running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying team-shape?

CallGet help
Cite this page

team-shape skill. Agent Mart. https://agentmart.dev/skills/oneiron-dev-team-shape (last updated 2026-10-24)