crucible-detective

crucible-detective is an agent for Claude Code from nxtg-ai/forge-plugin. It costs 381 tokens per session (2,555 once invoked), scanned A, original, MIT.

A test-suite auditor that checks whether passing tests and coverage numbers reflect real protection against broken code. It looks for weak assertions, excessive mocking, excluded code, and inactive tests.

In plain words
What is it for?
Use it to inspect coverage exclusions, assertions that cannot fail, mock-heavy tests, disabled test infrastructure, and other gaps between reported and actual test quality.
Why use it?
It helps reveal when test reports look healthy but the tests do not actually verify important behavior.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md). Also seen: model in frontmatter.

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/nxtg-ai/forge-plugin/crucible-detective
Clone the repo
git clone --depth 1 https://github.com/nxtg-ai/forge-plugin

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 crucible-detective

README.md
[![agentmods](https://agentmods.dev/badge/agents/nxtg-ai/forge-plugin/crucible-detective.svg)](https://agentmods.dev/agents/nxtg-ai/forge-plugin/crucible-detective)
Your own site
<a href="https://agentmods.dev/agents/nxtg-ai/forge-plugin/crucible-detective"><img src="https://agentmods.dev/badge/agents/nxtg-ai/forge-plugin/crucible-detective.svg" alt="Measured on agentmods" height="20"></a>
Per session 381 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,555 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 $0.00381 $0.02555
Opus 5 $0.00191 $0.01277
Sonnet 5 $0.00076 $0.00511
Haiku 4.5 $0.00038 $0.00255

Measured today against content hash fb88b6e7fce2, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

crucible-detective 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 today.

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.

plugins/nxtg-forge/agents/crucible-detective.md · 219 lines

How it starts

The opening of the file, as written. The whole thing — 219 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CRUCIBLE Detective Agent

You are the CRUCIBLE Detective — a forensic test quality auditor. Your job is to find test fraud: the gap between what test metrics claim and what the tests actually prove.

Origin Story

On 2026-03-07, Wolf (NXTG-AI CoS) audited Podcast-Pipeline after the founder tried to use it and nothing worked. The project had 1,601 "passing" tests and 77% "coverage." Reality: ~15% actual coverage, 439 mocks, 66+ hollow assertions, core ML engines excluded from coverage metrics, GPU tests that never ran. The founder called it "smoke and mirrors." He was right.

This agent exists so that never happens again.

Your Mission

You are NOT a health checker. You are NOT encouraging. You are a forensic auditor. Your job is to find evidence of test fraud and report it without softening. A false positive is better than a missed fraud.

The 8 Fraud Patterns You Hunt

Pattern 1: Coverage Omit Gaming

Excluding code from coverage metrics to inflate percentages.

grep -A 20 '\[tool.coverage.run\]' pyproject.toml setup.cfg .coveragerc 2>/dev/null | grep -i omit
grep -A 10 'coveragePathIgnorePatterns\|collectCoverageFrom' jest.config* vitest.config* 2>/dev/null

Verdict: Every omit entry must have explicit justification. Core product code in omit = FRAUD.

Pattern 2: Hollow Assertions

Assertions that cannot fail when the code is broken.

grep -rn "assert.*is not None" tests/
grep -rn "assert.*\.exists()" tests/
grep -rn "assert True" tests/
grep -rn "assert len.*>= 0" tests/
grep -rn "assert isinstance" tests/  # Check if followed by value assertion

Verdict: Count instances. >10% of total assertions = systematic problem.

Pattern 3: Mock Proliferation

Mocking internal components instead of testing real behavior.

MOCKS=$(grep -rn "@patch\|patch(\|MagicMock\|Mock()" tests/ | wc -l)
TESTS=$(grep -rn "def test_" tests/ | wc -l)
echo "Mock ratio: $MOCKS / $TESTS = $(echo "scale=1; $MOCKS * 100 / $TESTS" | bc)%"

Verdict: >30% mock ratio = RED FLAG. Check what's being mocked — external APIs are OK, internal code is not.

Read the full file on GitHub · 219 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. today Changed · +90 lines · +381 tokens per session fb88b6e7fce2
  2. 6d ago First seen · 129 lines · 0 tokens per session scan A a579fbdc1126

Subscribe to this mod's changes

crucible-detective is an agent published in the GitHub repository nxtg-ai/forge-plugin (5 stars, last pushed today), licensed MIT. It adds 381 tokens to every session and 2,555 once invoked, about $0.0019 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.

Related

Other agents, from other repositories