codex-game-studio/openspec/changes/expand-eval-framework-coverage/tasks.md

1. OpenSpec Planning

• 1.1 Create proposal for first-pass eval coverage expansion
• 1.2 Create design documenting CCGS/Truthmark-inspired decisions
• 1.3 Create specs for coverage and framework validation requirements
• 1.4 Validate OpenSpec change with openspec validate expand-eval-framework-coverage --strict

2. TDD Coverage

• 2.1 Restore/add tests/performance-evaluation-framework.test.ts
• 2.2 Add failing threshold tests for at least 30 scenarios, 10 workflows, 10 skills, and 6 roles
• 2.3 Add validation expectations for first-pass coverage checks
• 2.4 Run focused test and confirm RED before implementation

3. Eval Framework Data

• 3.1 Add role behavior rubric
• 3.2 Expand catalog to cover workflow prompt targets
• 3.3 Expand catalog to cover skill-maintenance, QA, review, readiness, and gate targets
• 3.4 Expand catalog to cover representative role prompt targets
• 3.5 Add scenario JSON and prompt markdown for each target

4. Validation Integration

• 4.1 Add coverage summary helper to src/performance-evaluation.ts
• 4.2 Add validation check for first-pass coverage thresholds
• 4.3 Keep manual-only validation free of real agent/judge calls

5. Verification

• 5.1 Run focused performance eval framework tests
• 5.2 Run npm run typecheck
• 5.3 Stage new tests before full suite
• 5.4 Run npm run test
• 5.5 Run npm run validate
• 5.6 Run OpenSpec validation and status
• 5.7 Run Truthmark check/index and verify no deleted tests