Test Suite Auditor — Gaps, Flakiness, False Confidence

Audit this test suite. The question is not "is there enough coverage" but "which of these tests would actually fail if the code broke" — and which are theater. Test code: ```{{language}} {{test_code}} ``` Code under test (for gap analysis): ```{{language}} {{production_code}} ``` Audit dimensions: 1. **Tests that cannot fail meaningfully** — assertion-free tests (call it and hope), assertions on constants, tests asserting the mock returns what the mock was told to return, snapshot tests nobody would read on failure. 2. **Mock suffocation** — tests where so much is mocked that the test re-states the implementation. Litmus: could the production code have an inverted condition or dropped call and this test still pass? Trace it, do not guess. 3. **Flakiness time bombs** — real time/dates (what happens when this runs at 23:59, on Feb 29, in a different TZ?), order-dependent tests sharing state, sleeps instead of synchronization, real network/filesystem, iteration over unordered collections with ordered assertions. 4. **Coverage gaps that matter** — map the production code's branches and failure paths (error handling, empty inputs, boundaries) against the tests. List UNTESTED behaviors ranked by blast radius, not by line coverage percent. Ignore trivial gaps (getters). 5. **Duplication** — clusters of tests that fail together for the same cause; propose which to keep. Output: - A table: test name → verdict (SOLID / THEATER / FLAKY-RISK / REDUNDANT) → one-line reason. - Top 5 fixes ranked by (confidence gained / effort), each with concrete rewritten test code. - Top 5 missing tests, each with a skeleton (name + arrange/act/assert outline). - One paragraph: overall, what does this suite actually protect, and what does it merely appear to protect?