claudeers.
// Claude Skills

octopus-skill

🐙 One brain, many arms — a curated prompt library for long-horizon agent work, compiled to each host (Claude Code · grok · Cursor · Codex). Arms: loop-graph…

Actively maintained
100/100
last commit 9 days ago
last release 21 days ago
releases 5
open issues 1
// star history

Install with your AI

Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.

Install and set up octopus-skill (claude-skill project) into my current project.
Found on https://claudeers.com/octopus-skill
Repo: https://github.com/levi-qiao/octopus-skill
Homepage/docs: https://github.com/levi-qiao/loop-skill
Detected install method: claude-skill → # copy this skill into .claude/skills/octopus-skill/
Category: skills. Platforms: cli.
Read the repo's README for exact setup and env vars, then install it and wire it into my project.

Claudeers Health Verdict:
active; community-verified: false. Confirm the source before running anything.
// or install directly (claude-skill)
# copy the skill dir into your project:
# .claude/skills/octopus-skill/   (or ~/.claude/skills/octopus-skill/ for all projects)
// or clone
git clone https://github.com/levi-qiao/octopus-skill

// compatibility

Platformscli
Operating systems
AI compatibilityclaude
LicenseMIT
Pricingopen-source
LanguageShell

octopus-skill 🐙

One brain, many arms. A curated library of battle-tested prompts for long-horizon agent work, compiled to whatever host you run — Claude Code, grok, Cursor, Codex. The methodology is shared; each arm adapts it to a host's native shape (a loop, or a goal). Like an octopus, one nervous system reaching into different environments and changing color to fit each one.

English · 简体中文

the graph: an executor node and a clean-context supervisor node communicating only through durable files

Why this exists

A strong model authoring a tuned, opinionated prompt beats hand-driving a host ad hoc — for repeatable, long-horizon, high-stakes work. (For a quick one-off, just type the task; the library's ROI is in reuse.) The durable value isn't any one mechanism — it's the discipline: "done" means verified not written, no test theater, no speculative building, forced convergence against growth, and hard owner red lines. That discipline is host-agnostic. octopus is where it lives, once, and gets compiled down to each host.

The arms

ArmUse whenShips
loop-graphyou'll drive it with /loop (Claude Code, grok, Cursor, shell); multi-round work that scope-creeps or fakes "done"; multi-milestone phases; executor and supervisor split across hosts/models; owner-gatedan executor node (works a single-source-of-truth ledger) + a clean-context supervisor node that re-verifies from outside and corrects via a one-way directives file — two loops
questyou'll hand it to a goal command that self-drives to done (grok /goal, a Codex task); a single self-contained goal, executor+reviewer in the same runone objective prompt that folds the discipline in and rides the host's own verifier — no second loop

Not sure which? The decision rule lives at the top of each arm's SKILL.md, and the host capability matrix is in lib/host-dialects.md.

The brain (lib/)

  • methodology.mdwhy each rule exists, tied to the specific failure mode of long agent runs it prevents. Read this to adapt rules without breaking them.
  • host-dialects.md — the single owner of per-host differences: loop/goal invocation syntax, adaptive-vs-interval behavior, and wake/notify/keep-alive primitives (grok Stop/Notification hooks, etc.).

Install

curl -fsSL https://raw.githubusercontent.com/levi-qiao/octopus-skill/main/install.sh | sh

Installs as a single /octopus skill for Claude Code and Codex — the umbrella routes to the right arm. Any existing /graphkit install is left untouched. To install from a local clone instead, run ./install.sh from the repo root.

Governance — keep it a library, not a junk drawer

octopus applies its own anti-bloat rule to itself: no prompt enters the library without a real consumer — a run it was actually proven on. Curated and opinionated beats comprehensive. Same bar the executor holds inside a run.

Credits

The loop-graph arm grew from real runs and community input — it began life as the standalone graphkit skill, and this repo is that project, evolved (the old graphkit URL redirects here). Special thanks to @BrightProgrammer7 — the migrate-blob-storage worked example and the design discussions that sharpened the milestone-gate and node/edge vocabulary.

License

See LICENSE.

// faq

What is octopus-skill?

🐙 One brain, many arms — a curated prompt library for long-horizon agent work, compiled to each host (Claude Code · grok · Cursor · Codex). Arms: loop-graph (executor + clean-context supervisor) · quest (one objective riding the host's own verifier).. It is open-source on GitHub.

Is octopus-skill free to use?

octopus-skill is open-source under the MIT license, so it is free to use.

What category does octopus-skill belong to?

octopus-skill is listed under skills in the Claudeers registry of Claude-compatible tools.

1 views
57 stars
unclaimed
updated about 1 month ago

// embed badge

octopus-skill on Claudeers
[![Claudeers](https://claudeers.com/api/badge/octopus-skill.svg)](https://claudeers.com/octopus-skill)

// retro hit counter

octopus-skill hit counter
[![Hits](https://claudeers.com/api/counter/octopus-skill.svg)](https://claudeers.com/octopus-skill)

// reviews

// guestbook

0/500

// related in Claude Skills

🔓

An agentic skills framework & software development methodology that works.

// skillsobra/Shell272,506MIT[ claude ]
🔓

Public repository for Agent Skills

// skillsanthropics/Python169,406[ claude ]
🔓

💫 Toolkit to help you get started with Spec-Driven Development

// skillsgithub/Python129,208MIT[ claude ]
🔓

AI coding assistant skill (Claude Code, Codex, OpenCode, Cursor, Gemini CLI, and more). Turn any folder of code, SQL schemas, R scripts, shell scripts, docs,…

// skillsGraphify-Labs/Python106,387MIT[ claude ]
→ see how octopus-skill connects across the ecosystem