Harden Albilich research harness and scheduler #1
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: phase2-soundness | |
| on: | |
| push: | |
| pull_request: | |
| jobs: | |
| unit-and-adversarial: | |
| runs-on: ubuntu-latest | |
| timeout-minutes: 30 | |
| steps: | |
| - uses: actions/checkout@v4 | |
| with: | |
| fetch-depth: 0 | |
| - uses: actions/setup-python@v5 | |
| with: | |
| python-version: "3.12" | |
| - uses: actions/setup-node@v4 | |
| with: | |
| node-version: "22" | |
| - name: Install the runtime-attested Codex CLI | |
| run: npm install --global @openai/codex@0.152.0 | |
| - name: Attest the real Codex executable boundary | |
| run: python -m agents.generation.phase2.backend_contract codex --binary codex | |
| - name: Compile | |
| run: python -m compileall -q agents/generation/phase2 | |
| - name: Enforce append-only scheduler policy epochs | |
| if: github.event_name == 'pull_request' || github.event.before != '0000000000000000000000000000000000000000' | |
| env: | |
| POLICY_EPOCH_BASE: ${{ github.event.pull_request.base.sha || github.event.before }} | |
| run: python -m agents.generation.phase2.policy_epoch_guard --base-ref "$POLICY_EPOCH_BASE" | |
| - name: Phase 2 unit, migration, replay, and adversarial tests | |
| env: | |
| PYTHONDONTWRITEBYTECODE: "1" | |
| run: python -m unittest discover -v -s agents/generation/tests -p 'test_phase2*.py' | |
| - name: Audit experiment archive labels and checksums | |
| run: python -m agents.generation.phase2.experiment_audit | |
| - name: Validate checked-in JSON files | |
| run: python -m json.tool agents/generation/phase2/schemas/phase2_patch.schema.json >/dev/null |