Arena , make N sub-agents compete in a tournament until one answer survives
When Claude keeps giving a bad answer, spin up N sub-agents (default 100, --quick for 16) with different strategy cards, run a single-elimination bracket where they attack and defend each other's solutions, and let a rubric-scored judge crown one survivor
- What
- When Claude keeps giving a bad answer, spin up N sub-agents (default 100, --quick for 16) with different strategy cards, run a single-elimination bracket where they attack and defend each other's solutions, and let a rubric-scored judge crown one survivor
- Cost
- Free
- Needs
- an agentic coding setup that can spawn sub-agents (Claude Code); Python 3 for bracket.py.
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor: for when Claude keeps giving a bad answer: instead of retrying, make 100 versions of it fight to the death. It spins up N sub-agents on the exact same task, each with a different strategy card (reasoning mode, workflow, strategy), runs a single-elimination bracket where they attack, defend and revise each other's solutions, and a judge scores every match on a written rubric until one answer survives. Ships with bracket.py for all the bookkeeping. by @Jakeschincariol, listed here with credit to its creator. Honest caveats: token-hungry by design, the default 100 agents means about 595 sub-agent calls, so use --quick (16 agents, 91 calls) for everyday use and confirm before launching a full run; the orchestrator never competes and never judges. Skill Harbor never reviews the code, review it yourself before use. Discovered via GitHub.
Version:
Quick answer
Arena is a free Muse skill for when Muse keeps giving a bad answer: it spins up N sub-agents with different strategy cards and lets them compete until one answer survives. Each sub-agent works on the same task, then a single-elimination bracket has them attack and defend each other's solutions, and a rubric-scored judge crowns one survivor.
Install
Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 11, 2026📌 Pinned version: commit df07b8ePrerequisites: an agentic coding setup that can spawn sub-agents (Claude Code); Python 3 for bracket.py. Install "Arena , make N sub-agents compete in a tournament until one answer survives" for me. It gives my agent @Jakeschincariol's tournament workflow: N sub-agents get the same task with different strategy cards, a single-elimination bracket with attack/defend/revise rounds, rubric-scored judging, all bookkeeping through bracket.py. Note: token-hungry by design , default 100 agents (~595 sub-agent calls); use --quick (16 agents) for everyday use. Repository: https://raw.githubusercontent.com/Jakeschincariol/arena-skill/df07b8e518e10cdecd8cd7712d9958b6f396250e/skills/arena/SKILL.md Pinned version: install the code at commit df07b8e518e10cdecd8cd7712d9958b6f396250e exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Fetch the SKILL.md file (and any helper files, including bracket.py) from the repository path into a temporary folder 2. Safety check: review the SKILL.md and scripts for anything suspicious (unexpected network calls, shell commands, credential harvesting). This repo should contain zero secrets in code, credentials only via the secure vault, allowed hosts declared in the SKILL.md. Verify that holds here; STOP on any red flag and tell me. 3. Install it as a skill: copy SKILL.md and its helper files into the agent's skills directory, in a folder named "arena". 4. Verify with no network calls: frontmatter valid, files in place. 5. Report what was installed, where, and what I still need to do myself (e.g. confirm Python 3 is available, decide on a default arena size, and always confirm with me before launching a full 100-agent run , it is expensive). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.
Saved to your recent installs. Find it anytime on /connect.
🔒 Security, provenance and what this touches
- It can run commands on your Muse’s computer.
What the scan found — excerpts from the install prompt
- runs commands — Prérequis : un environnement de code agentique capable de lancer des sous-agents (Claude Code) ; Python 3 pour
The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Verification certificate
Dated facts about what was checked for this listing, and how. This is not a safety certification.
- Pinned version
- df07b8e · 2026-10-13
- Sandbox install test
- Install ran in our test sandbox · Nov 11, 2026
- Creator identity
- Not verified
- Fingerprint watch
- Last checked · Nov 19, 2026 : no change detected in the files linked by this install prompt
- Install prompt analysis
- 2026-10-13 : runs commands
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Questions
How do I install a skill?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.