test-auditlisted

# Test Audit Systematic audit of test files against the project's test conventions and the user's codified test quality standards. ## Review mindset Approach each test as potentially wrong — not as a passing artifact to skim: - **Don't assume correctness.** A test that exists and passes is not assumed correct. Every test is under audit — evaluate whether it actually proves what it claims. - **Review every test individually.** Not per-file patterns — per-test. Each `it()` block gets its own evaluation against the audit dimensions below. - **Default to thorough.** No shortcuts, no skimming, no "the rest look fine." If a test's reliability is in question, mutation-test it. - **Follow fable-mode discipline.** Write a stage map before starting. Verify each stage with a check that can fail. Self-critique before delivering findings. ## Before starting Load these sources fresh: 1. **Project test conventions**: Read `AGENTS.md` (or `CLAUDE.md`) from the project root. Focus on the test conventions section. 2. **Code standards + preference recall**: Discover the standards notes first — the set grows and hardcoded lists go stale: `vault_search({ query: "code standards", filters: { tags: ["code-standards"], type: "reference", properties: { lifecycle: "living" } } })` Then `vault_read_note` the audit-relevant results (currently the testing note), plus any newer note matching this audit's focus or the repo's stack. Then recall the dated evidence trail for t