agentleFS
imMamdouhaboammarimMamdouhaboammar/anti-slop-engineering/skills/mandela-guard/SKILL.md

mandela-guard

Skill4 starsChanged 31 days agoMITView on GitHub

Audit test suites, benchmarks, and evals for the 8 Mandela Leakage Patterns where tests confirm themselves, verifiers judge their own designs, or assertions are tautological.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Zero Swallowed AssertionErrors: Never catch AssertionError in tests unless specifically testing an assertion library via an inverted test …

    line 27

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Mandela Test-Oracle & Leakage Guard

Eliminate test-oracle self-grading, tautological assertions, and circular mock confirmations.

Why This Skill Is Needed & Target State

The "Mandela Effect" in LLM testing occurs when a test suite passes not because the code works, but because the test grades itself. Common failure modes:

  1. Tautological Assertions: assert True, expect(true).toBe(true), or self-comparisons expect(val).toBe(val).
  2. Circular Mock Confirmation: Mocking a return value mock.return_value = 'x' and immediately asserting assert mock() == 'x' without exercising real domain logic.
  3. Swallowed Test Failures: try: ... except AssertionError: pass silently turning test breakages into passes.
  4. Frame-Injected Evals: Prompt benchmarks where the prompt tells the evaluator to confirm success.

The target state is independent ground truth: every test assertion must be capable of failing when the implementation is broken (A/B/A falsification).

Invariant Rules

  1. External Ground Truth: Assertions must evaluate computed outputs against independent expected constants, not self-derived values.
  2. Zero Swallowed AssertionErrors: Never catch AssertionError in tests unless specifically testing an assertion library via an inverted test runner.
  3. No Circular Mocking: Mock only at system boundaries (I/O, network, clocks). Never mock the function under test.

CLI Usage

# Scan test directory for Mandela leakage patterns
bunx anti-slop-engineering mandela tests/ --json

# Run in strict mode (fails on warnings)
bunx anti-slop-engineering mandela --strict

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.