Installs into .claude/skills of the current project.
Are you the author of Vector Store Grahama1970 Agent Skills?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/majiayu000-vector-store-grahama1970-agent-skills-claude-skill-registry)
---
name: vector-store
description: Transient vector store service for fast similarity search (FAISS)
triggers:
- vector store
- faiss
- similarity search
- vector index
allowed-tools: curl
commands:
- name: serve
description: Start the vector store service
usage: ./run.sh serve
- name: sanity
description: Run sanity tests
usage: ./sanity.sh
metadata:
port: 8600
endpoints:
- /index
- /search
- /reset
- /info
provides:
- vector-store
composes:
- memory
- embedding
- edge-verifier
- task-monitor
---
# Vector Store Skill
Transient vector store service for fast similarity search using FAISS (IndexFlatIP).
Designed to be used by other skills (memory, edge-verifier) to accelerate KNN search.
## Configuration
- `VECTOR_STORE_PORT`: Port to listen on (default: 8600)
## API
### POST /index
Add vectors to the index.
```json
{
"ids": ["id1", "id2"],
"vectors": [[0.1, 0.2, ...], [0.3, 0.4, ...]],
"reset": false
}
```
### POST /search
Search for nearest neighbors.
```json
{
"query": [0.1, 0.2, ...],
"k": 10
}
```
Returns:
```json
{
"ids": ["id2", "id1"],
"scores": [0.95, 0.88]
}
```
### DELETE /reset
Clear the index.
### GET /info
Get index stats.