Skip to content
Automation
Skill

/skill-prototype

Prototype one risky assumption within a fixed budget, then keep or discard the result

From plugin
octo
4.1k70 skills49 agents53 commands18 hooks
Install
$ npx -y skills add nyldn/claude-octopus --skill skill-prototype --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/skill-prototype

Context preview

The summary Claude sees to decide when to auto-load this skill.

Prototype one risky assumption within a fixed budget, then keep or discard the result

SKILL.md

skill-prototype.SKILL.md
name: skill-prototype
description: "Prototype one risky assumption within a fixed budget, then keep or discard the result"
disable-model-invocation: true

> **Host: Codex CLI** — This skill was designed for Claude Code and adapted for Codex. > Cross-reference commands use installed skill names in Codex rather than `/octo:*` slash commands. > Use the active Codex shell and subagent tools. Do not claim a provider, model, or host subagent is available until the current session exposes it. > For host tool equivalents, see `skills/blocks/codex-host-adapter.md`.

Bounded prototype

Use a prototype to answer one named question, not to start implementation by a different name.

Contract

Agree on the question, hypothesis, deadline, artifact path, and success signal before writing. Record the source revision. Store artifacts through `scripts/plan-storage.sh`, outside the installed plugin cache.

Native read-only plan mode permits a proposal only. Wait for an explicit execution handoff before creating files. Choosing prototype mode does not authorize deployment, provider calls, browser login, repository rewrites, or new permissions.

Stop at the deadline. Return observations, a `keep`, `discard`, or `inconclusive` verdict, and the disposition of each artifact. Automatic cleanup may remove only artifacts created by this run after ownership checks. Retain anything the user changed.

The completion record contains `question`, `hypothesis`, `deadline`, `artifact_path`, `source_revision`, `observations`, `verdict`, and `disposition`.

Adapted from `prototype` in `mattpocock/skills` at commit `3cca18b368ae95cdbdebbff572ccafa662551015` under the MIT License. See `THIRD_PARTY_NOTICES.md`.

Read more
Ships withocto

Every AI model has blind spots. Claude Octopus supports twelve external provider integrations — Codex, Antigravity CLI, Copilot, Qwen, Ollama, Perplexity, OpenRouter, OrcaRouter, OpenCode, Cursor CLI, Grok, and Kimi Code — alongside the built-in Claude Code

Get the whole plugin

Other skills on octo.