cardio

cardio is a command for Claude Code from a777ance/claude-code-homelab. It costs 64 tokens per session (926 once invoked), scanned A, original, MIT.

A command that asks several independent in-harness reviewers to consider the same question and returns the majority opinion. It is intended for quick judgment calls without an API key or API spending.

In plain words
What is it for?
Use it for one-off design decisions, identifying a likely bug, or checking a factual judgment. Repeatable tasks with labelled examples are directed to a separate statistical jury tool.
Why use it?
It reduces the chance that one uncertain or unusual answer determines the result.

Command for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: mentions subagents.

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 commands/a777ance/claude-code-homelab/cardio
Clone the repo
git clone --depth 1 https://github.com/a777ance/claude-code-homelab

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for cardio

README.md
[![agentmods](https://agentmods.dev/badge/commands/a777ance/claude-code-homelab/cardio.svg)](https://agentmods.dev/commands/a777ance/claude-code-homelab/cardio)
Your own site
<a href="https://agentmods.dev/commands/a777ance/claude-code-homelab/cardio"><img src="https://agentmods.dev/badge/commands/a777ance/claude-code-homelab/cardio.svg" alt="Measured on agentmods" height="20"></a>
Per session 64 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 926 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 5.1 $0.00064 $0.00926
Opus 5 $0.00032 $0.00463
Sonnet 5 $0.00013 $0.00185
Haiku 4.5 $0.00006 $0.00093

Measured 5d ago against content hash 530ee0f570f4, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

cardio 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 5d 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.

.claude/commands/cardio.md · 63 lines

How it starts

The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Run the self-consistency sampling doctrine inside this harness: don't consume a single warm draw where a verdict matters — empanel a jury, then let a plurality vote outvote the idiosyncratic draws.

Question to deliberate: $ARGUMENTS

First: is this the right tool?

  • One-off judgment call (a design decision, "which of these is the bug", a factual call you need now) → empanel the in-harness jury below.
  • Measurable or repeatable task (you have or can build a labelled set, or you'll ask this shape of question often) → the statistical, calibrated jury tool (jury_claude.py, with calibrate/form/diet/strength) lives in the localDNS repo under 04-user-services/ai-orchestration/jury-claude/. Its Dirichlet stopping rule and measured jury size beat the fixed fan-out below. Come back here only for the un-repeatable calls.

Empanel the jury (concurrent vote)

  1. Empanel 5 jurors at once, each with a different framing. In a single message, spawn 5 concurrent juror subagents (Task tool, subagent_type: "juror"), each handed the same question ($ARGUMENTS) but a different answer-preserving framing directive, so the draws decorrelate by construction rather than by luck (identical prompts to a locked-decoding model collapse into one reasoning path). Juror 1 gets no directive (plain); jurors 2–5 get one each:

    • skeptic — work one step at a time; before accepting each step, ask how it could be wrong.
    • restate — restate the question and list what's given vs. asked; derive from those facts only.
    • cross-check — solve, then reach the answer a second way (estimate / work backwards) and reconcile.
    • avoid-the-trap — name the most tempting wrong approach first, then deliberately avoid it.

    The framing changes the approach, never the question or the answer. Ask each to end with a canonical ANSWER: line.

  2. Tally. Collect each juror's ANSWER: line, normalize (lowercase, strip surrounding punctuation/whitespace), and count. The plurality is the working verdict.

Read the full file on GitHub · 63 lines

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. 5d ago First seen · 63 lines · 64 tokens per session scan A 530ee0f570f4

Subscribe to this mod's changes

cardio is a command published in the GitHub repository a777ance/claude-code-homelab (2 stars, last pushed 25d ago), licensed MIT. It adds 64 tokens to every session and 926 once invoked, about $0.0003 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.