[Anionix]

Summary

• keep generated robustness rows consistent with recorded parameters
• normalize serialized robustness enums and preserve legacy claims reportability
• reject recursive evidence imports and no-op truncation recipes
• distinguish read rejection from verification VALUE_MISMATCH
• resolve relative bounded run directories and restore the full lifecycle contract comment

Addresses the contract/evidence findings tracked in #176 (historical review findings #109-#113 and #126). Native fuzz follow-ups remain separate.

Verification

• nix develop --command ruff check src tests
• nix develop --command pyright src/format_bench/model.py src/format_bench/robustness/evidence.py src/format_bench/robustness/mutations.py src/format_bench/robustness/paths.py src/format_bench/robustness/runner.py
• nix develop --command env PYTHONPATH=src .venv/bin/pytest -q tests/test_model.py tests/test_robustness_cases.py tests/test_robustness_artifacts.py tests/test_robustness_targets.py tests/test_robustness_profile.py tests/test_report.py tests/test_release.py

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3676d78481

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".