search_tools_respond

When to use

Call search_tools_respond only after search_tools has returned a response with "type": "clarification". This tool is never called on its own — it always follows a prior search.

Input schema

query_id················································································stringrequired

The unique query identifier returned in the clarification response from search_tools.

answers················································································Record<string, string>required

Key-value map answering the clarification questions. Keys must match the "key" field from each question.

Examples

Answering clarification questions

{
  "query_id": "q_abc123def456",
  "answers": {
    "language": "Python",
    "deployment": "cloud"
  }
}

// After search_tools returns clarification questions,
// Claude automatically calls search_tools_respond:
// {
//   "query_id": "q_abc123def456",
//   "answers": {
//     "language": "Python",
//     "deployment": "cloud"
//   }
// }

// Cursor handles the clarification flow automatically.
// When search_tools asks for clarification, Cursor
// prompts the user and calls search_tools_respond
// with the collected answers.

Response format

This tool always returns a result response — no further clarification is requested. The response structure is identical to the result format from search_tools.

{
  "type": "result",
  "recommendation": {
    "name": "Pinecone",
    "description": "Managed vector database for ML embeddings at scale",
    "category": "database",
    "language": "Python",
    "license": "proprietary",
    "health_tier": "healthy",
    "confidence": 0.91,
    "reasons": [
      "Cloud-native managed service — no infrastructure to manage",
      "First-class Python SDK with async support",
      "Optimized for high-dimensional embedding vectors"
    ]
  },
  "alternatives": [
    {
      "name": "Weaviate",
      "confidence": 0.85,
      "description": "Open-source vector database with built-in vectorization"
    },
    {
      "name": "Chroma",
      "confidence": 0.82,
      "description": "Embedded vector DB with first-class developer ergonomics"
    }
  ]
}

Error codes

400················································································Bad Requestoptional

Invalid query_id or malformed answers object. Ensure the query_id matches a prior clarification response.

404················································································Not Foundoptional

The query_id has expired or does not exist. Start a new search with search_tools.

500················································································Internal Erroroptional

Pipeline processing error. An unexpected failure occurred while completing the search.

Related tools

• search_tools — Start a new tool search (prerequisite for this tool)
• report_outcome — Report usage feedback after using a recommended tool