Follow up to: #3582504 _**NOTE: Before working on this issue, verify it's still relevant and makes sense.**_ **Description:** Build an observability layer around context selection so administrators and developers can understand what was selected, why it was selected, and whether it helped. This also creates the foundation for future experimentation. **Includes:** - `3.1` Context effectiveness feedback - Full `3.2` Selection audit trail - `3.3` Context A/B testing **Scope:** - Record selection reasoning and relevant selection metadata. - Add UI or reports for reviewing selected context. - Add a feedback mechanism for context effectiveness. - Explore A/B testing support for different context strategies or items. - Avoid storing sensitive raw context unless explicitly needed and safe. **Acceptance criteria:** - Selection decisions can be inspected after an agent run. - Feedback can be captured and associated with context items or runs. - Stored audit data is bounded, cache-safe, and privacy-conscious. - A/B testing, if added, is opt-in and deterministic. [x] AI Assisted Issue