- Home
- Skills
- Testing & Debugging
- Test Quality Auditor
Works with the AI Agents you already use
Test Quality Auditor
Audit test suites for false confidence, flakiness, and hidden CI suppressions to ensure they catch real regressions.
$5
Test Quality Auditor
Example session with this skill installed
Audit the quality of the tests in src/auth/service.spec.ts. Are these tests actually going to catch a bug in the login flow?
- Read your context and instructions
- Compiled the test quality auditor
I have audited src/auth/service.spec.ts and the associated CI configuration. While the coverage is technically 92%, the suite has significant structural weaknesses that would allow authentication bypasses to reach production.
Test Quality Review — src/auth/service.spec.ts
Summary
The current suite provides false confidence. Most tests assert that the mock functions were called rather than verifying the resulting state or token validity, meaning the logic could be completely broken while the tests remain green.
False confidence
[src/auth/service.spec.ts:45] Mocking the verifyPassword function to always return true within the test setup. This would still pass if the actual database comparison logic was deleted.
Fix: Use a test database instance or a narrower mock that requires a valid hash match.
[src/auth/service.spec.ts:89] Tautological assertion: expect(result).toBeDefined(). This passes even if the service returns an error object instead of a valid session.
Fix: Assert on specific properties of the session object, such as result.userId and a valid JWT structure.
Flaky / Blind spot
[src/auth/service.spec.ts:112] Blind spot: No tests exist for the account lockout mechanism after five failed attempts. A regression in the rate-limiter would not be caught.
[src/auth/service.spec.ts:150] Flaky: Test uses new Date() to check token expiration. This will fail intermittently in CI depending on execution lag.
Fix: Use a fixed seed or mock the system clock.
Not enforced
[.github/workflows/ci.yml:22] The auth test job uses continue-on-error: true. Failures in this specific suite do not currently block the merge pipeline.
Next steps
- Remove
continue-on-errorfrom the CI workflow to ensure failures are blockers. - Refactor the
verifyPasswordmock to validate input against a set of known test hashes. - Replace generic
toBeDefined()checks with schema-based assertions for the session response.
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
High test coverage is often a vanity metric. AI-generated tests frequently pass by asserting trivial truths or mocking the very logic they are supposed to verify, leaving real regressions undetected.
What it does
- Identifies tautological assertions where tests pass regardless of underlying logic changes.
- Detects over-mocking that prevents tests from verifying actual side effects or returns.
- Finds flaky patterns like unseeded randomness, wall-clock time dependencies, and sleep-based waits.
- Audits CI configurations for 'continue-on-error' flags or skipped tests that mask failures.
Why this beats prompting it yourself
General prompts usually focus on writing new tests or fixing syntax errors. This skill uses a specific trust rubric and scanning workflow to find structural weaknesses that a standard LLM would typically overlook or even introduce themselves.
Use cases
- Reviewing PR test suites to ensure new code is actually protected.
- Auditing legacy test suites that have high coverage but frequent production bugs.
- Validating agent-generated tests to ensure they aren't just asserting 'true is true'.
Known limitations
Does not perform a stylistic review of test code or naming conventions unless they cause reliability issues. Requires access to test files and CI configuration files.
Does not perform stylistic reviews of test code. Requires access to test files and CI configuration files to identify hidden suppressions.
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 any payment.
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
3 uses
Times used to date
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 months ago