test(operator_commands): audit + replace low-value smoke tests with output assertions

[rysweet]

Summary

Test-quality audit of the Simard suite. The CI coverage metric
(cargo llvm-cov --workspace --lib --bins --ignore-filename-regex 'tests?/')
already reports 82.84% line coverage — well above the 70% target — so the
useful work here is the quality half of the mandate: finding tests that
execute code without verifying behavior and turning them into real
behavior assertions, rather than padding the metric.

The audit surfaced a coherent cluster of low-value tests in the
operator-command output layer: the renderers print directly via println!, so
their tests could only assert "does not panic" while discarding the actual
rendered text. That left the real behavior unverified — status filtering,
priority/title/slug sorting, count lines, <none> placeholders,
terminal-control sanitization, and concise-label formatting.

Changes

Behavior-preserving refactor: extract pure helpers that return the rendered
lines, and have the existing print_* wrappers forward them to stdout
(rendered output is byte-identical). Then rewrite the smoke tests to assert
exact output.

• operator_commands::format — added format_labeled + *_section_lines
  helpers; converted 9 *_does_not_panic tests into output assertions
  (sanitization strips ANSI escapes, Items→Item singularization, goal
  filtering + priority sort, empty <none> placeholders, terminal-handoff
  Some/None branches incl. the missing-last_output_line fallback).
• operator_commands::goals — added render_lines on the goal register
  view/section; converted 2 print_does_not_panic tests into full-output
  assertions (per-section counts, <none>, concise-label rendering).
• operator_commands_terminal::commands — terminal_recipe_list_does_not_panic
  discarded its Result (a probe can return Err without panicking); now
  asserts the probe succeeds and the descriptor list is non-empty and every
  descriptor carries a name.

Net: 12 assertion-free / discarded-result smoke tests replaced with
behavior-verifying assertions. The thin print_* wrappers are still
exercised once each so the println! paths stay covered.

Audit methodology

Scanned all ~9,600 #[test]/#[tokio::test] functions for anti-patterns
(assert!(true), assert_eq!(x, x), assertion-free bodies, discarded
Results). Most heuristic hits were false positives (helper-based assertions,
e.g. overseer::signal's has() helper). The genuine low-value cluster was
the does_not_panic stdout smoke tests fixed here. No tautological
(assert!(true) / assert_eq!(x, x)) tests exist in the tree.

Evidence

• Coverage (criterion: >70%) — cargo +nightly-2026-07-01 llvm-cov --workspace --lib --bins --ignore-filename-regex 'tests?/' --json --summary-only: overall 82.84% lines (143676/173432); functions 83.43%,
  regions 83.90%. operator_commands module 80.9% before this change.
• Tests — all 32 tests in the three touched modules pass; the strengthened
  assertions verify exact rendered output (see files).
• fmt — cargo fmt --all -- --check clean.
• clippy — cargo clippy --lib --tests --all-features --locked -- -D warnings
  clean; pre-commit cargo clippy --release --no-deps -- -D warnings and
  pre-push cargo clippy --all-targets --all-features --locked -- -D warnings
  both passed.
• Diff focused — 3 files, all in the operator-command rendering layer; no
  unrelated edits.

Surfaces changed

Internal only. The print_* functions render byte-identical stdout (the
*_lines helpers are the single source of truth the wrappers now use), so
there is no user-facing behavior change and no docs update is required.

[github-actions]

📊 Coverage Summary

Generated by cargo llvm-cov --workspace --summary-only (nightly, excluding test files)

Module	Lines	Covered	Coverage
Total	175449	145072	82.7%

_{Coverage data from CI run. Test files matching tests?/ are excluded from line counts.}

[rysweet]

CI update: all 16 checks green (pre-commit/verify, coverage, install-real, e2e-dashboard, cargo-deny/audit/vet, npm-audit, GitGuardian). The CI coverage job independently confirms the metric: Total 145072/175449 lines = 82.7%, verifiably above the 70% target. No coverage regression from the audit — the 12 reworked tests assert byte-identical rendered output, and the thin print_* wrappers stay exercised.