Add --last-failed / --lf flag

[MatthewMckee4]

Summary

Closes #464.

• Add --last-failed / --lf CLI flag to re-run only tests that failed in the previous run
• Workers now write per-run failed_tests.json files alongside existing stats/diagnostics/durations
• Aggregated failed test names are persisted to last-failed.json at the cache root after each run
• When --last-failed is passed, the partition step filters collected tests against the last-failed list
• If no previous failures exist (or no cache), all tests run normally

Test plan

• last_failed_reruns_only_failures -- verifies only failing tests re-run
• last_failed_lf_alias -- verifies the --lf alias works
• last_failed_with_no_previous_failures_runs_all -- all pass on previous run, so all re-run
• last_failed_without_previous_run_runs_all -- no cache at all, all tests run
• Full test suite passes (681/681)
• CLI docs regenerated
• Pre-commit checks pass

🤖 Generated with Claude Code

[codspeed-hq]

Merging this PR will not alter performance

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

✅ 1 untouched benchmark

---

_{Comparing add-last-failed-flag (beba6e2) with main (777db22)}