ingest-feedback

To convert unstructured client feedback into testable QA rules, paste the feedback document into the chat. The system extracts discrete items, maps each to a rule for traceability, and appends them to qa-rules.md while saving a dated feedback log.

The best way to track repeated issues in feedback logs is by scanning prior feedback logs and existing qa-rules.md entries. This mechanism detects recurring items, enabling you to trigger a QA-loop regression audit to verify whether previously flagged problems were addressed.

Feedback traceability works by mapping each extracted item from pasted feedback to a structured rule in the qa-rules.md file. The full feedback document and its rule mappings are simultaneously saved to a dated markdown log at feedback/YYYY-MM-DD.md.

Yes, you need to rely on the qa-rules schema to extract QA rules from review notes. The system reads existing references and maps each unstructured observation to a structured rule, ensuring the newly appended rules conform to the repository's QA system.

Yes, you can trigger a QA-loop regression audit when feedback repeats. The system scans history to detect repeated issues by comparing new feedback against prior logs, enabling a regression audit to verify all items were addressed during the next QA loop.

A limitation of logging feedback to dated markdown files is that each session saves to a single feedback/YYYY-MM-DD.md path, meaning multiple ingests on the same date may overwrite prior logs unless managed externally, and manual review is still required to validate rule extraction.