Summary

Adds semantic hard-fails for artifacts that are structurally complete but ineffective.

Why

The repo now has interactive capabilities and routing. This PR prevents regressions where an artifact looks complete but does not state the work it removes, does not map evidence to consequence, or declares interaction without real controls plus reusable output.

Changes

• Adds hard-fails for missing effectiveness job, missing evidence-to-consequence mapping, and ineffective interactive capabilities.
• Adds tests/artifacts/bad/complete-but-ineffective.html.
• Updates tests/artifacts/good/interactive-tool.html so its evidence has a consequence/action link.
• Confirms merged interactive templates still pass audit-artifact.py --min-score 90.

Validation

• Bad fixture blocks as expected.
• scripts/validate-golden-artifacts.sh
• scripts/validate-runtime.sh
• Individual audit of merged templates under templates/.

Rollback

Revert this PR to remove the stricter semantic hard-fails and fixture changes.