Add tests for #1493

[prompt-driven-github]

Summary

Adds tests based on the requirements in #1493.

Test Files

• tests/test_checkup_negative_ac.py

Test Coverage

• Total Tests: 14
• Framework: pytest + unittest.mock
• Status: passing (14/14)
• Test Plan Coverage: 14/14 planned cases implemented (100%)

What These Tests Verify

These tests verify that pdd checkup correctly detects missing negative/adversarial acceptance-criteria coverage:

1. Checkup fails when negative AC not tested (3 tests) — issue_aligned=false → success=False, no push when adversarial coverage is missing
2. Checkup passes when negative AC is covered (2 tests) — issue_aligned=true → success=True when adversarial near-match cases (gpt-4, openai/gpt-4o, azure/gpt-4.1-mini) are covered
3. Happy-path tests alone do not satisfy alignment (2 tests) — Step 5 passing + Step 7 issue_aligned=false → orchestrator still fails; passing tests alone are not sufficient
4. Adversarial probe language in prompts (3 tests) — Static content assertions that agentic_checkup_step7_verify_LLM.prompt references issue_aligned and checkup_review_loop.py contains adversarial probe family instructions
5. Near-match pricing regression fixture (4 tests) — End-to-end regression guard for the original Add output-inclusive token and cost estimation primitives #1357/chore: PDD sync for #1357 #1361 failure: gpt-4 must not borrow pricing from near-match models

Contract Test Summary

N/A - no OpenAPI spec found

Accessibility Audit Summary

N/A - not a web test

Manual Testing Summary

N/A - all tests are automated and deterministic with no live LLM credentials required

Test Execution

python -m pytest tests/test_checkup_negative_ac.py -q

Next Steps

1. Review the generated tests
2. Run tests locally to verify
3. Adjust tests if needed
4. Mark PR as ready for review

Closes #1493

---

Generated by PDD agentic test workflow (18-step)