circular-validation-audit

circular-validation-audit is a skill for Claude Code, Codex from yogsoth-ai/stress-test. It costs 90 tokens per session (1,325 once invoked), scanned A, a copy of circular-validation-audit, Apache-2.0.

A pre-build check for a validator, such as a simulation or benchmark, that is meant to test a theory. It looks for circular reasoning, where the test is built from the same assumptions it claims to verify.

In plain words
What is it for?
Use it before building tests, sandboxes, simulations, or benchmarks for a theory or recovery method.
Why use it?
It prevents a validator from producing convincing but meaningless results. A passing result has little value if success was guaranteed by the validator's design.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. 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 skills/yogsoth-ai/stress-test/circular-validation-audit
Any agent
npx skills add yogsoth-ai/stress-test --skill circular-validation-audit
Clone the repo
git clone --depth 1 https://github.com/yogsoth-ai/stress-test

Made for: Claude Code, Codex.

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 circular-validation-audit

README.md
[![agentmods](https://agentmods.dev/badge/skills/yogsoth-ai/stress-test/circular-validation-audit.svg)](https://agentmods.dev/skills/yogsoth-ai/stress-test/circular-validation-audit)
Your own site
<a href="https://agentmods.dev/skills/yogsoth-ai/stress-test/circular-validation-audit"><img src="https://agentmods.dev/badge/skills/yogsoth-ai/stress-test/circular-validation-audit.svg" alt="Measured on agentmods" height="20"></a>
Per session 90 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,325 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin 100% copy Near-identical to another mod 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 $0.00090 $0.01325
Opus 5 $0.00045 $0.00662
Sonnet 5 $0.00018 $0.00265
Haiku 4.5 $0.00009 $0.00133

Measured 6d ago against content hash e1ab3bd41b13, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

circular-validation-audit 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 6d 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.

Origin

This is a copy

100% identical to circular-validation-audit — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.

skills/circular-validation-audit/SKILL.md · 56 lines

How it starts

The opening of the file, as written. The whole thing — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Circular Validation Audit

A pre-build gate for any validator we are about to construct — a sandbox, simulation, or benchmark that is supposed to verify a theory. The catastrophic failure mode: we build the validator using the theory's own assumptions as the validator's ground-truth machinery, run the theory's method on the validator's output, and celebrate when it "recovers" the truth — but the recovery was guaranteed by construction, not by the theory being right. The PASS carries zero evidential weight. This is the single most expensive mistake available, because a circular validator can burn enormous compute producing confident, meaningless confirmations. So this audit runs FIRST, on the validator DESIGN, before a line of it is built.

The core risk, stated precisely

A validator V tests a theory T by: generating data D from a ground-truth generator G, running T's recovery method R on D, and checking if R(D) matches G's hidden truth. The test is non-circular only if G is NOT itself an instance of T's central assumptions. If G is built from the same functional factorization / the same noise law / the same mechanism that T claims to discover, then "R recovers G" is a tautology: we encoded the answer into the question. The audit's job is to find every place where G secretly contains T.

Non-circularity matrix (the central artifact)

Rows = the theory's load-bearing claims/assumptions (the ones the validator is meant to test). Columns = the validator's ground-truth construction choices (how each mechanism is implemented, what functional forms or noise laws the generator uses, what intervention semantics the validator actually applies).

Each cell asks: does this construction choice ASSUME this claim?

validator: construction choice 1 validator: construction choice 2 validator: construction choice 3
claim: load-bearing claim A ? ? ? ?
claim: load-bearing claim B ? ? ? ?
claim: load-bearing claim C ? ? ? ?
? ? ? ?

Read the full file on GitHub · 56 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. 6d ago First seen · 56 lines · 90 tokens per session scan A e1ab3bd41b13

Subscribe to this mod's changes

circular-validation-audit is a skill published in the GitHub repository yogsoth-ai/stress-test (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 90 tokens to every session and 1,325 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. It is 100% identical to circular-validation-audit, differing in 0 lines, and is treated as a copy.