# Build Azure AI Search Apps with TypeScript

Azure AI Search integrations require careful SDK configuration across indexing, authentication, and multiple query modes. This skill provides focused TypeScript patterns for building and reviewing those workflows.

## Install

```bash
npx skillstore add sickn33/azure-search-documents-ts
```

## Metadata

- Status: approved
- Slug: sickn33-azure-search-documents-ts
- Skillstore revision: r2
- Version status: missing
- Tree hash: 7eb79f070a4c6c5a3ff1c13a9b7fdf420b8078e4f6e34858fbbb6999c01052fd
- Author: sickn33
- GitHub username: sickn33
- License: MIT
- Repository: https://github.com/sickn33/antigravity-awesome-skills/tree/main/skills/azure-search-documents-ts
- Ref: f9e2c34b4f19c7f3e6b0a1e93227b5f77cc12526
- 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, env\_access
- Quality score: 78
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/sickn33-azure-search-documents-ts
- Manifest: https://skillstore.pages.dev/api/skills/sickn33-azure-search-documents-ts/manifest

## Capabilities

- Configures SearchClient and SearchIndexClient with DefaultAzureCredential and environment settings.
- Defines an Azure AI Search index with text, filterable, facetable, and vector fields.
- Demonstrates document uploads and mixed batch upload, merge, and delete actions.
- Shows full-text, vector, hybrid, and semantic search request patterns.
- Provides filtering, faceting, autocomplete, and suggestion examples.
- Lists common SDK types and production-oriented query guidance.

## Use Cases

- Create a Search Prototype: Build a TypeScript prototype that indexes product records and supports text, vector, and hybrid queries.
- Design Search Relevance: Compare semantic ranking, vector retrieval, filters, and facets for a search experience.
- Review Azure Integration: Check SDK authentication, index definitions, batch operations, and query options before implementation.

## Prompt Templates

### Set Up a Client

```
Create a minimal TypeScript setup for Azure AI Search using DefaultAzureCredential. Use environment variables for the endpoint and index name.
```

### Define and Populate an Index

```
Design a product index with searchable text, category facets, and a 1536-dimension vector field. Include a batch upload workflow.
```

### Add Hybrid Search

```
Build a hybrid query that combines user text with an embedding. Return ten results and include selected document fields.
```

### Plan a Production Search Flow

```
Review my Azure AI Search requirements. Propose semantic ranking, filters, facets, batching, validation, retry handling, and tests with clear assumptions.
```

## Limitations

- Examples contain placeholders and require an existing Azure AI Search service, index permissions, and local package installation.
- The vector examples depend on an external embedding function that the skill does not implement.
- The snippets omit comprehensive error handling, input validation, retries, logging controls, and automated tests.
- API availability and type names can vary by installed SDK version and Azure service configuration.

## Best Practices

- Prefer DefaultAzureCredential and least-privilege identities instead of embedding administrator keys in source code.
- Batch document changes and use incremental updates to reduce indexing overhead.
- Validate field names, filter values, vector dimensions, service capabilities, and query performance in the target environment.

## Anti Patterns

- Do not treat example placeholders, omitted types, or the undefined embedding function as production-ready code.
- Do not build authorization filters directly from untrusted client input.
- Do not request total counts, semantic features, or large nearest-neighbor sets without measuring cost and latency.

## Security Audit

- Audited at: 2026-07-23T22:42:48.881\+00:00
- Summary: All 45 static findings are false positives. The command detections are Markdown fences, TypeScript template literals, or inline code, while the URL and environment access are normal configuration examples. No prompt injection, credential exfiltration, or hidden malicious intent was found in SKILL.md.

## Stats

- Views: 86
- Downloads: 11
- Favorites: 0
- Popularity score: 0