| name | map-context-and-link-evidence |
| description | Use this to answer "what do we know about this?" with a Memgraph-first context map and link relevant Qdrant memories as graph evidence. |
Context Map And Memory Linking
Use when asked for durable cross-entity context.
Context Map Format
## Context Map: <entity or topic>
### Entity
- Slug: <slug>
- Labels: <graph labels>
- Domain: <real | fictional | speculative>
- Status: <active | archived | other>
### Relationships
- <relationship> -> <connected entity> (role: <role>, kind: <kind>)
### Linked Memories
- [<kind>] <memory summary> (agent: <agent>, created: <date>)
### Nextcloud References
- <path> - <brief description>
### Summary
<2-3 sentence synthesis>
Linking Rules
- Memgraph first for relationships
- Qdrant exact lookup second when Memgraph has
MemoryEntry.qdrant_id: run python3 qdrant/get_memory.py "<point-id>"
- Qdrant MCP semantic search only when graph traversal is sparse or the entity is unknown
- shared Nextcloud
/Desk/ cues and Nextcloud references third for supporting docs
- when a Qdrant memory deserves graph structure, represent it as linked
MemoryEntry structure with slug qdrant:<point_id>
- include
nc_refs when a memory points to a Nextcloud artifact
- after successful Memgraph linking, annotate the Qdrant point with
python3 qdrant/set_graph_link.py
General Archivist Return Format
Use this when the task is curation work but not primarily a context-map response:
## Handoff Complete
**Task:** <brief restatement>
**Status:** <complete | partial - needs X | blocked - needs Y>
### Deliverables
- Memgraph: <nodes, edges, schema or query updates>
- Qdrant: <memories stored, linked, groomed>
- Nextcloud: <paths updated, if any>
### For the user
<concise explanation of what durable context was added or clarified>
### Follow-up needed
<which agent should do what next>