# Build AgentDB Semantic Search

Teams need relevant document retrieval for RAG and knowledge bases. This skill guides AgentDB setup, embedding storage, indexing, API search, and relevance tuning.

## Install

```bash
npx skillstore add dnyoussef/agentdb-semantic-vector-search
```

## Metadata

- Status: approved
- Slug: dnyoussef-agentdb-semantic-vector-search
- Version: 1.0.0
- Author version: 1.0.0
- Skillstore revision: r1
- Version status: valid
- Tree hash: 2a92395801d3febde693446f85dd207116f3af1c41bce255d0c29472a9a42189
- Author: DNYoussef
- GitHub username: DNYoussef
- License: MIT
- Repository: https://github.com/DNYoussef/ai-chrome-extension/tree/main/.claude/skills/agentdb/when-building-semantic-search-use-agentdb-vector-search
- Ref: 72d5025b022c77f7a51bdf5c1637c689c80e89d1
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network
- Quality score: 80
- Quality tier: silver
- Public page: https://skillstore.pages.dev/skills/dnyoussef-agentdb-semantic-vector-search
- Manifest: https://skillstore.pages.dev/api/skills/dnyoussef-agentdb-semantic-vector-search/manifest

## Capabilities

- Guides AgentDB initialization with vector dimensions and HNSW indexing.
- Shows how to embed documents and store metadata with vectors.
- Outlines REST API endpoints for semantic search queries.
- Describes metadata filtering, hybrid retrieval, and result re-ranking.
- Defines success metrics for relevance, latency, and top result quality.

## Use Cases

- Build RAG Retrieval Backend: Create a vector search layer that returns relevant context for LLM responses.
- Improve Knowledge Base Search: Add semantic matching so users can find documents by meaning, not exact keywords.
- Evaluate Search Relevance: Plan retrieval metrics, top result checks, and tuning cycles for search quality.

## Prompt Templates

### Plan a Basic Index

```
Use this skill to plan an AgentDB semantic search index for my document corpus. Include dimensions, metadata fields, and success metrics.
```

### Design Document Ingestion

```
Use this skill to design a document ingestion flow with embeddings, metadata storage, and duplicate handling for AgentDB.
```

### Create Search API Plan

```
Use this skill to outline REST API endpoints for semantic search, filtering, pagination, and result ranking.
```

### Optimize Hybrid Retrieval

```
Use this skill to improve an existing AgentDB search system with hybrid retrieval, re-ranking, and relevance evaluation.
```

## Limitations

- Requires an available AgentDB package or compatible vector search implementation.
- Requires access to an embedding model and related credentials.
- Examples need adaptation for authentication, validation, and production deployment.
- Does not include full operational guidance for scaling or monitoring.

## Best Practices

- Validate embedding dimensions before storing vectors.
- Keep metadata filters simple and aligned with user search needs.
- Measure relevance with representative queries before production rollout.

## Anti Patterns

- Do not assume keyword search alone will satisfy semantic retrieval goals.
- Do not store document text without clear privacy and retention controls.
- Do not tune latency without checking result relevance.

## Security Audit

- Audited at: 2026-07-06T12:04:08.991\+00:00
- Summary: The flagged reconnaissance items are false positives caused by normal hybrid search documentation. The external command finding is a Markdown TypeScript code fence, and the network finding is a documentation URL. No semantic prompt injection, data exfiltration intent, or malicious behavior was found in the reviewed files.

## Stats

- Views: 220
- Downloads: 8
- Favorites: 0
- Popularity score: 0