golang-test-reviewlisted

# /golang-test-review A focused Go-test-engineering review. Run it over the changed `*_test.go` files (and any test harness / fixtures / test tooling) **after** writing tests and before merge. It does not replace `code-reviewer` — it adds the test-specific lens that general review misses. ## How to run 1. Identify the changed test surface: `git diff --stat <base>...HEAD -- '*_test.go' 'go/test/**' 'go/cmd/testlatency/**'`. 2. Read each changed test file and score it against the checklist below. 3. Report findings by severity (CRITICAL / HIGH / MEDIUM / LOW), each with `file:line`, the rule violated, and a concrete fix. End with a verdict: APPROVE (no CRITICAL/HIGH) / CHANGES-REQUESTED. ## Checklist ### 1. Behavior over surface (the highest-signal rule) - Does the test pin the **invariant under change**, or just re-walk lines for coverage? A test that would still pass if the behavior broke is a no-op — flag it CRITICAL. (AGENTS.md Rule 9.) - Are assertions specific (exact value / error) rather than "not nil" / "len > 0" where an exact check is possible? - Negative + edge cases present, not only the happy path? ### 2. Determinism (non-negotiable) - No dependence on ambient state: live repo, `$HOME`, real clock, network, `.evolve/runs/`. Tests build isolated state (`t.TempDir()` + `git init`, or the shared `fixtures.NewWorkspace`). - No real `time.Sleep` for synchronization where an injected clock / channel / poll-with-deadline would be deterministic. Rea