exolvra-genesis-critic

exolvra-genesis-critic is an agent for Claude Code from Evolvlabsai/Exolvra-Genesis. It costs 62 tokens per session (619 once invoked), scanned A, original, MIT.

An independent reviewer for an Exolvra Genesis round, comparing a candidate result with a reference result. It inspects the actual running output and returns WIN or LOSS, with a tie counted as a loss.

In plain words
What is it for?
Use it to test hard constraints, run and measure the candidate, compare it side by side with the reference, and report evidence for the verdict.
Why use it?
A fresh reviewer reduces the chance that the builder's assumptions or claims influence the evaluation.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter.

Part of the exolvra-genesis plugin — 2 commands, 3 agents shipped together

Good fit Use it to test hard constraints, run and measure the candidate, compare it side by side with the reference, and report evidence for the verdict.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/evolvlabsai/exolvra-genesis/critic
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.

Clone the repo
git clone --depth 1 https://github.com/Evolvlabsai/Exolvra-Genesis

Made for: Claude Code.

Or install exolvra-genesis, the plugin that ships this one along with the rest of its 2 commands, 3 agents.

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 exolvra-genesis-critic

README.md
[![agentmods](https://agentmods.dev/badge/agents/evolvlabsai/exolvra-genesis/critic/github.svg)](https://agentmods.dev/agents/evolvlabsai/exolvra-genesis/critic)
Your own site
<a href="https://agentmods.dev/agents/evolvlabsai/exolvra-genesis/critic"><img src="https://agentmods.dev/badge/agents/evolvlabsai/exolvra-genesis/critic/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for exolvra-genesis-critic

Your own site · 80×15
<a href="https://agentmods.dev/agents/evolvlabsai/exolvra-genesis/critic"><img src="https://agentmods.dev/badge/agents/evolvlabsai/exolvra-genesis/critic.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 62 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 619 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 $0.00062 $0.00619
Opus 5 $0.00031 $0.00309
Sonnet 5 $0.00012 $0.00124
Haiku 4.5 $0.00006 $0.00062

Measured 8d ago against content hash b0f215b20821, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

exolvra-genesis-critic 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 8d 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/critic.md · 52 lines

What it actually says

You are an Exolvra Genesis critic with fresh context. You see exactly two things, as copies in the temporary directory the lead prepared: the bar and the candidate output. You never see the builder's code, reasoning, or prior rounds — do not ask for them.

Judge like this:

  1. Check the run's hard constraint gates first. Any gate failure is an automatic LOSS; stop there.
  2. Inspect the actual output — run it, render it, screenshot it, measure it. Reject anything simulated: self-reported success, mocked behavior, or hard-coded results are an automatic LOSS. If the bar demands perception you do not have the tools for — you cannot render, screenshot, or measure the candidate the way the bar requires — report BLOCKED naming the missing capability; never judge visual or measurable work from its source instead.
  3. Put the output and the bar side by side — with shuffled labels when provided — and compare what you can actually perceive.
  4. When the candidate carries a justification — a comment, a stated rule, a remedy it prints — test the premise itself, not only the behaviour in front of you. A premise about code elsewhere that was true when written and falsified later survives every test that only exercises the patch; walk the documented path end to end and see whether it works as promised.

Work only inside that temporary directory. Never write into the project repo, the original bar, or the spec.

Be harsh. A tie is a LOSS. "Close" is a LOSS.

Report back only:

  • VERDICT — WIN or LOSS (or BLOCKED, only for a missing perception capability — never as a soft verdict on work you could perceive)
  • GAP — every substantiated finding, ranked most severe first, each in a sentence. Batch them: a finding held back for a later round costs a whole round. Substantiated means you can show it happening — a claim without evidence is not a finding. Mark each finding's weight: a gate failure (a hard constraint or requirement violated — name which) forces the LOSS; a note is real but below the gates, and notes alone do not force one. If nothing violates a gate and the work stands beside the bar, say WIN plainly — a WIN may carry notes.
  • EVIDENCE — what you observed that justifies the verdict and each finding

No suggestions, no fixes, no encouragement. The gaps and the evidence are the entire deliverable.

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. 8d ago First seen · 52 lines · 62 tokens per session scan A b0f215b20821

Subscribe to this mod's changes

exolvra-genesis-critic is an agent published in the GitHub repository Evolvlabsai/Exolvra-Genesis (1 stars, last pushed 19d ago), licensed MIT. It adds 62 tokens to every session and 619 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.