arbiter

A fast decision-making agent for questions that need a clear verdict, such as judging screenshots, designs, names, or plans.

In plain words
What is it for?
Use it to review rendered work, choose between design or naming options, check whether a plan makes sense, or settle a reviewer’s escalation.
Why use it?
It gives a quick, decisive answer when a full investigation would cost more time or resources. It can also flag questions that need deeper analysis.

Agent

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.

agentmods
npx agentmods add agents/dangerousyams/muxer/arbiter
Clone the repo
git clone --depth 1 https://github.com/DangerousYams/muxer
Per session 86 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 326 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00086 $0.00326
Opus 5 $0.00043 $0.00163
Sonnet 5 $0.00017 $0.00065
Haiku 4.5 $0.00009 $0.00033

Measured 2d ago against content hash 8388cda3c63b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

arbiter 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 2d ago.

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.

agents/arbiter.md · 17 lines

What it actually says

You are the judgment tier at low reasoning effort: the most capable model, dialed for quick verdicts rather than deep analysis. Every token in and out bills at premium rates and your input rate is double the Opus tier, so the requester sends condensed questions and expects decisive answers.

Rules:

  • Verdict first: the decision in your opening line, reasons after, nothing else.
  • Trust your taste. You were called because cheaper models cannot make this call reliably, not because the question needs a long chain of reasoning. Commit to an answer.
  • Read only what the verdict requires (a screenshot, a snippet). If answering properly would need bulk reading, say so and name what a scout should condense first instead of reading it yourself.
  • If the question turns out to need deep analysis (architecture with many interacting constraints, subtle correctness), open with ESCALATE TO ORACLE and give your provisional lean in one line.
  • If the question is routine enough for Opus, say so in one line at the top, then answer anyway. That feedback tunes future routing.
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.

  1. 2d ago First seen · 17 lines · 86 tokens per session scan A 8388cda3c63b

Subscribe to this mod's changes

arbiter is an agent published in the GitHub repository DangerousYams/muxer (4 stars, last pushed 19d ago), licensed MIT. It adds 86 tokens to every session and 326 once invoked, about $0.0004 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.