A test-writing approach that checks stated requirements without reading the code being tested. It focuses on observable results, such as returned values, files, and exit codes.

It is for writing tests from requirements or a requirements document, including boundary cases and behaviours that must be refused.

Tests based on the implementation can pass even when the requested behaviour is wrong. This approach keeps tests tied to what the software must do and exposes unclear requirements.

crux-tests — Agent for Claude Code by mehdiforoozandeh/crux

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/mehdiforoozandeh/crux

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.

README.md

[![agentmods](https://agentmods.dev/badge/agents/mehdiforoozandeh/crux/crux-tests/github.svg)](https://agentmods.dev/agents/mehdiforoozandeh/crux/crux-tests)

Your own site

<a href="https://agentmods.dev/agents/mehdiforoozandeh/crux/crux-tests"><img src="https://agentmods.dev/badge/agents/mehdiforoozandeh/crux/crux-tests/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.

Your own site · 80×15

<a href="https://agentmods.dev/agents/mehdiforoozandeh/crux/crux-tests"><img src="https://agentmods.dev/badge/agents/mehdiforoozandeh/crux/crux-tests.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 80 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 457 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.

Model	Per session	Once invoked
Fable 5.1	$0.00080	$0.00457
Opus 5.5	$0.00032	$0.00183
Sonnet 5.5	$0.00016	$0.00091
Haiku 4.5	$0.00008	$0.00046

Measured 6d ago against content hash e432252a0238, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

crux-tests 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.

skills/evolve-crux/agents/crux-tests.md · 43 lines

What it actually says

crux-tests — test the requirement, not the code

When invoked

Read the requirement or RD. Restate, in one line each, what must be true when it works.
For each, write the assert that would fail if it were not true. Name the assert after the behaviour, not the function.
Include the boundaries the requirement states: the empty case, the "must refuse" case, and any case the requirement explicitly allows.
Say which requirements you could not test, and what is missing.

Rules

Never read the implementation. If you cannot write the test without it, the requirement is under-specified — report that, which is the more useful finding.
A test that mirrors the code is not a test. It survives any refactor and notices no wrong answer.
Assert the observable: return values, files written, exit codes.

Output

The tests, plus the requirements you could not express as one and what is missing.

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.

6d ago First seen · 43 lines · 80 tokens per session scan A e432252a0238

Subscribe to this mod's changes

crux-tests is an agent published in the GitHub repository mehdiforoozandeh/crux (17 stars, last pushed today), licensed MIT. It adds 80 tokens to every session and 457 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-11-01.

Claude Code Testing agent-skills ai-agents claude cockpit hypothesis-testing

crux-tests next to 6 agents from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
crux-testsmehdiforoozandeh/crux	Agent	17	today	MIT	A	80	Claude Code
doc-reviewershinpr/nautilus	Agent	7	16d ago	MIT	A	33	Claude Code, Cursor
hypothesis-verifiershinpr/nautilus	Agent	7	16d ago	MIT	A	32	Claude Code, Cursor
prototype-generatorshinpr/nautilus	Agent	7	16d ago	MIT	A	35	Claude Code, Cursor
codebase-analyzershinpr/nautilus	Agent	7	16d ago	MIT	A	25	Claude Code, Cursor
knowledge-distillershinpr/nautilus	Agent	7	16d ago	MIT	A	31	Claude Code, Cursor
hypothesis-verifiershinpr/claude-code-discover	Agent	12	2mo ago	MIT	A	42	Claude Code