v0.2 — Run Feedback and Factor Scoring Updates

Quantinal v0.2 introduces human run feedback comments and improves factor-based score explainability across the API and console workflows.

Highlights

Human Feedback on Runs

• Added support for run-level human comments via nested run endpoints.
• Comments can be added, updated, or cleared for a run and are returned in run payloads.
• Feedback comments are now integrated into agent context flows:
  • Included in previous-run self-awareness context.
  • Included in tracker history enrichment for owned runs.

New Run Comment Endpoint

• Added:
  • PATCH /v1/api/trackers/{tracker_id}/runs/{run_id}
• Behavior:
  • Trims comment input.
  • Empty string clears comment (null).
  • Owner-only access and visibility rules enforced.

Factor Scoring Update

• Expanded documentation for factor outputs in run results:
  • baseline_score
  • score_adjustment_factors
• Clarified deterministic scoring flow:
  • baseline is engine-injected (previous run score or 50 for first run),
  • factor impacts adjust the baseline before quality compression.

Notes

• This release is backward compatible for existing run trigger/list/get/delete flows.
• Run comments are updated through PATCH /v1/api/trackers/{tracker_id}/runs/{run_id}.