Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
DeepSeek
Muse+22 moreClaude CodeCursorCodex CLIDeepSeekMuse
Dots
Manus
Grok
Grok Bot+18 agentsHonest Tests
Stop AI coding agents from writing fake tests. An agent skill + CLI that detects tests that can't fail, and proves it with mutation testing.
Free
honest-tests
Example session with this skill installed
Audit the tests in src/checkout and tell me which ones can't actually fail.
- Read your context and instructions
- Compiled the honest-tests
Test honesty audit: src/checkout
10 of 14 tests can't fail. Mutation score for pricing.ts: 15.4% (22 of 26 mutants survived).
Confirmed errors
| Test | Rule | Why it can't fail | Fix |
|---|---|---|---|
| pricing.test.ts:39 "computes the discounted price" | swallowed-assertion | the catch eats the AssertionError, and 70 is actually wrong | drop the try/catch, expect 80 |
| checkout.test.ts:12 "qualifies members" | asserts-on-mock | asserts the value the mock was told to return | call the real checkoutTotal |
Surviving mutants
- pricing.ts:22
>=→>: no test checks the exact free-shipping threshold (50).
Connects securely to your tools. The creator never sees your data.
About this skill
Your AI wrote 200 tests. Many of them can't fail.
AI-written test suites fail in predictable ways: they assert that a value equals itself, assert on the return value of a mock they just configured, mock the very module under test, wrap assertions in try/catch, or never assert at all. They stay green forever and protect nothing.
honest-tests teaches your agent to write tests that go red when the behavior they name breaks, and to prove it.
Two modes
- Write tests: eight hard rules (never weaken an assertion to pass, expected values come from outside the code under test, mock only boundaries...), then a self-audit before reporting done.
- Audit tests: a phased procedure: detect with the open-source CLI, verify every finding by reading the test, prove weak tests with mutation testing, and report a headline like "10 of 14 tests can't fail" with a fix for each.
What it catches (13 rules): tests with no assertions, tautologies (expect(x).toBe(x)), asserting on mocks or on literals the test just built, mocking the unit under test, swallowed or catch-only assertions, matchers that never run, snapshot-only tests, conditional assertions, and leftover .only/.skip.
Mutation testing makes small deliberate bugs in your source (>= → >, true → false, return x → undefined) on a temporary copy and shows which ones no test notices: concrete, line-level proof of what your suite doesn't check.
Runs locally. MIT-licensed, open source: github.com/mintwork/honest-tests.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.
JavaScript and TypeScript tests only (vitest, jest); pytest support is planned. Mutation testing runs one mutant at a time, so large files take a while; cap runs with --max-mutants. Static rules are heuristics: the skill's Phase 2 verifies each finding before reporting it.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed today
- Free with an Agensi account
- Software
Node.js 22+JS/TS project (vitest or jest)@mintwork/honest-tests
Permissions
Allowed Hosts
File Scopes
Runs locally. No data leaves your machine apart from downloading the open-source CLI from npm.