fix(automation): kill uncatchable run_cargo mutants

[martin-kolinek]

Problem

The scheduled-exhaustive CI leg (anvil-mutants-full) failed with 2 missed mutants in crates/automation/src/lib.rs (run_cargo):

• crates/automation/src/lib.rs:76 — replace run_cargo -> Result<(), AppError> with Ok(())
• crates/automation/src/lib.rs:83 — delete ! in run_cargo

Root cause

duct::cmd(...).run() already returns Err on a non-zero exit status (duct's default). That made the if !output.status.success() block effectively unreachable dead code — so flipping ! or stubbing the body with Ok(()) was unobservable, hence uncatchable. As a bonus bug, output.stdout/stderr were read in the bail message but are always empty because stdio is inherited (piped live).

Fix

• Add .unchecked() so duct no longer short-circuits, making the status check the single observable success/failure decision.
• Simplify the now-correct bail message (dropped the always-empty stdout/stderr).
• Add test_run_cargo_success and test_run_cargo_reports_failure; the failure test kills both mutants.

Why the diff-scoped PR check never caught this

anvil-mutants-diff only mutates lines in the PR diff (--in-diff). run_cargo was added 2026-03-18, ~3 months before cargo-mutants CI landed (#33, 2026-06-16), and no PR has touched those lines since — so they never appeared in a mutation diff. The scheduled full sweep is the safety net that surfaced it.

Verification

• File-scoped cargo mutants --file crates/automation/src/lib.rs: 4 caught, 1 unviable, 0 missed.
• Full-workspace cargo mutants --workspace run in progress; 0 missed at ~50% at time of opening. (Change is isolated to one file, so it cannot affect other crates.)
• cargo test -p automation, nightly fmt, and just spellcheck all clean.

[Codecov Comments Bot (codecov-commenter)]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.0%. Comparing base (829341e) to head (def6d23).

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #58   +/-   ##
=======================================
  Coverage   100.0%   100.0%           
=======================================
  Files          50       50           
  Lines        2652     2652           
=======================================
  Hits         2652     2652           

Flag	Coverage Δ
linux	?
linux-arm	?
scheduled	?
windows	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR fixes automation::run_cargo so that mutation testing can reliably observe (and therefore kill) mutants related to the command’s success/failure decision, and adds focused tests to prevent regressions.

Changes:

• Make run_cargo use duct::cmd(...).unchecked().run()? so non-zero exits don’t short-circuit into an error before the explicit status.success() check.
• Simplify the failure path to avoid reporting stdout/stderr that aren’t captured when running with inherited output.
• Add unit tests covering both a known-success invocation and a known-failure invocation to kill the previously missed mutants.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.