feat: map RunContract quality hints to five advisory dimensions

[devkade]

Summary

Map RunContract advisory quality hints into five explicit advisory dimensions: Alignment, Evidence, Completeness, Simplicity, and Risk.

Parent roadmap: #114
Track: A — RunContract Harness core
Tier: 1 — actionable now

Problem

RunContract currently exposes advisory quality hints, but the design notes call for a clearer five-dimensional quality vocabulary. Without explicit dimensions, future scoring/rendering work can drift into opaque aggregate scoring or overlap confusingly with existing workflow validation and quality-probe readiness checks.

Proposal

Represent RunContract quality as dimension-level advisory signals with reasons. Keep the result non-authoritative: workflow validation and evidence gates remain the completion authority.

Scope

• Add a RunContract quality dimension vocabulary:
  • alignment
  • evidence
  • completeness
  • simplicity
  • risk
• Represent each dimension as pass | warn | fail plus reason strings.
• Derive dimension signals from existing workflow state, evidence, artifacts, risks, blockers, and completion criteria.
• Preserve existing completion validation behavior.
• Add tests for representative ready/warn/fail cases.

Non-goals

• No numeric aggregate score.
• No hard approval/completion authority from quality hints.
• No tool-call blocking from score.
• No replacement of existing workflow validation.
• No replacement of src/domain/quality-probe.ts runtime readiness matrix.
• No GitHub/PR/kapi-agent/Ragna semantics in RunContract core.

Acceptance criteria

• RunContract quality exposes the five advisory dimensions with status and reasons.
• Missing evidence degrades the Evidence dimension without completing/failing the workflow by itself.
• Pending decisions and blockers are reflected in Completeness and/or Risk as advisory signals.
• Existing validation remains the source of completion authority.
• Tests cover at least one non-GitHub fixture.
• npm run verify passes.

Verification

• Run targeted RunContract tests.
• Run npm run verify.
• Search core RunContract code for adapter-specific terms.

References

• Kapi → Ilchul RunContract Harness roadmap #114
• feat: add soft quality scoring hints for RunContract status #100
• Obsidian design note: kapi-run-contract-refactor-loop-and-ragna-review.md