Test Coverage Auditor

The Problem

Your CI dashboard shows 78% line coverage. Your CTO calls it "good enough." But that 78% doesn't tell you that the payment processing module at src/payments/refund.ts has zero tests, that the auth token validator at src/auth/validateToken() has a test that only checks toBeDefined(), and that the CSV parser at src/utils/parse.ts has tests but none cover null input. The coverage number looks fine. The untested critical paths are a time bomb.

What You Get

• Detect your testing framework automatically — Jest, Vitest, Pytest, Go testing, Rust #[test], Mocha — so you get language-appropriate mapping rules without configuration
• Map every source file to its test files using framework-specific naming conventions (login.ts → login.test.ts, test_login.py, store_test.go) and flag files with zero test coverage
• Categorize untested exports by risk — Critical (auth, payments, data persistence), Warning (utilities, transformations), Info (constants, types) — so you write tests for what matters first
• Identify tested-but-incomplete coverage — files that have test files but miss null/empty/boundary/error-state edge cases, which static coverage tools report as "covered"
• Detect weak assertions — tests that pass without asserting meaningful outcomes (expect(result).toBeDefined(), bare function calls with no assert), inflating confidence without verifying behavior
• Generate a prioritized remediation plan — top 3 highest-risk untested functions, quick-win pure functions testable in under 5 minutes, and framework recommendations when no testing infrastructure exists

Why This Beats Prompting It Yourself

Running jest --coverage gives you a percentage and a per-file heat map. It doesn't tell you that the 3 untested files are all in your payments directory, that the "tested" file only exercises one of four exported functions, or that the test for your token validator uses a mock that bypasses the actual validation logic. This skill reads both source and test files, matches actual function calls against exports, and produces a risk-weighted gap map — the difference between "we have coverage" and "we have coverage of what matters."

Example Session

You: "Audit test coverage for this Node.js project before we start the testing sprint"

Agent:

TEST COVERAGE AUDIT REPORT
══════════════════════════

CRITICAL — PUBLIC API WITH ZERO TESTS
─────────────────────────────────────
File                       Function                Lines   Risk
─────────────────────────  ──────────────────────  ──────  ────────────────────
src/payments/refund.ts     refundOrder()           30-88   Handles money
src/payments/refund.ts     validateRefundAmount()  90-112  Financial validation
src/auth/validateToken.ts  verifyJWT()             15-42   Authentication

WARNING — TESTED BUT MISSING EDGE CASES
───────────────────────────────────────
File                Function    Missing Cases
──────────────────  ──────────  ────────────────────────────────────────
src/utils/parse.ts  parseCSV()  null input, empty string, malformed rows

TESTED BUT WEAK ASSERTIONS
──────────────────────────
File                  Test             Problem
────────────────────  ───────────────  ──────────────────────────────
tests/export.test.ts  "runs exporter"  No assertion on output content

TOP 3 PRIORITIES
────────────────
1. validateRefundAmount() — auth bypass risk on refund edge cases
2. verifyJWT() — current test mocks away actual verification
3. parseCSV() — null/empty input

Use Cases

• Planning a test-writing sprint — get a prioritized queue of what needs tests before you start
• Pre-deployment risk assessment — identify untested critical paths before go-live
• Inheriting an unknown codebase — map the test debt landscape in 5 minutes
• QA risk concentration analysis — find where untested code overlaps with money/auth paths
• Validating that coverage tool numbers match reality — pair with this audit to catch inflated metrics
• Team sprint planning — assign test-writing tasks based on risk priority, not file order

Known Limitations

This is static source-to-test mapping — it cannot execute tests or measure runtime coverage percentages. Go/Rust colocated tests require checking same-directory patterns, not just tests/ folders. Generated code, vendored dependencies, and ORM migrations may produce false "untested" flags if exclusions aren't configured. For runtime coverage numbers, pair with pytest --cov or jest --coverage.

Upgrade to Pro

Free maps the coverage gaps. Test Coverage Auditor Pro ($5) closes them: framework-aware test generation, edge-case matrices, mutation testing, and CI coverage trending. Upgrade when hand-writing tests is the bottleneck — Pro version.