refactor(bench): enforce GDC diagnostic measurement boundaries (#959)

[DecisionNerd]

Summary

• Add gdc_measurement_policy.py with fail-closed checks that live GDC evidence keeps certification: false, rejects misattributed BenchExec authority on in-process runners, and never treats resources.correctness_authority as a performance-gate input.
• Wire all four GDC suite adapters through the shared boundary enforcer.
• Register gdc-snb-bi in the benchmark measurement inventory.

Test plan

• PYTHONPATH=harness uv run --locked python -m unittest tests.test_gdc_measurement_policy
• PYTHONPATH=harness uv run --locked python -m unittest tests.test_gdc_snb_bi -k live
• python3 scripts/ci/benchmark-measurement-policy.py

Part of #959 (lifecycle/GDC measurement-boundary slice).

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository: CurateLabs/graphforge/.coderabbit.yaml

Review profile: CHILL

Plan: Advanced

Run ID: 2646fa98-42c3-4740-b03e-79e37ddaba2e

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

Warning

Billing warning: we have not been able to collect payment for this subscription for more than 72 hours. Please update the payment method or pay any pending invoices in Billing to avoid service interruption.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}