| name | zworkforce-rag-curation |
| description | Curate zWorkforce tenant-scoped memory and RAG behavior including memory records, tags, embeddings, local/Qdrant vector backends, reindexing, stale knowledge cleanup, privacy boundaries, and retrieval quality. |
zWorkforce RAG Curation
Keep tenant knowledge useful without crossing privacy or retention boundaries.
Workflow
- Identify tenant, agent scope, memory namespace, tags, source artifacts, and
retention policy.
- Remove stale, duplicate, or ungrounded memories only with the correct curator
authority and audit trail.
- Validate local feature-hash vectors or Qdrant/OpenAI-compatible embedding
configuration.
- Reindex after schema, embedding, source, or tag changes.
- Test retrieval quality with representative queries and cite source records.
References
docs/SECRET-MANAGEMENT.md
zworkforce/rag.py
zworkforce/db_tasks.py
tests/test_v3_rag_artifacts.py
README.md Memory / RAG section
Output
Report memory scope, changes made, sources, reindex status, retrieval checks,
privacy risk, and follow-up curation tasks.