AI & LLM

AI Agent Tool Registry JSON Example

A JSON example of an AI agent's tool registry / manifest — a list of callable tools with names, descriptions, and JSON Schema parameters. Copy-ready for function-calling agents and toolkits.

JSON

Generate mock data Open in Formatter

{
  "tools": [
    {
      "name": "search_orders",
      "description": "Search orders by customer email or order id.",
      "parameters": {
        "type": "object",
        "properties": {
          "query": {
            "type": "string",
            "description": "Email or order id"
          },
          "limit": {
            "type": "integer",
            "default": 10
          }
        },
        "required": [
          "query"
        ]
      },
      "returns": {
        "type": "array",
        "items": {
          "type": "object"
        }
      }
    },
    {
      "name": "issue_refund",
      "description": "Issue a refund for an order. Requires confirmation.",
      "parameters": {
        "type": "object",
        "properties": {
          "order_id": {
            "type": "string"
          },
          "amount": {
            "type": "number"
          }
        },
        "required": [
          "order_id"
        ]
      },
      "dangerous": true
    }
  ],
  "version": "2025-05-01"
}

Field Reference

toolsrequiredarray<object>The set of callable tools exposed to the agent/model

tools[].namerequiredstringUnique, snake_case tool name the model references when calling

tools[].descriptionrequiredstringPlain-language purpose — the model relies on this to choose the right tool

tools[].parametersrequiredobjectJSON Schema for the arguments; drives validation and prompting

tools[].dangerousoptionalbooleanCustom flag to gate destructive tools behind human confirmation

Variants

OpenAI tools arrayThe same registry shaped for the OpenAI tools parameter.

OpenAI tools array

Generate mock data Open in Formatter

{
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "search_orders",
        "description": "Search orders by customer email or order id.",
        "parameters": {
          "type": "object",
          "properties": {
            "query": {
              "type": "string"
            }
          },
          "required": [
            "query"
          ]
        }
      }
    }
  ]
}

Common Use Cases

→Declaring the toolset available to a function-calling agent
→Generating tool schemas dynamically from your backend functions
→Gating dangerous tools behind confirmation or permissions

ai agenttoolsfunction callingregistrymanifesttoolkit

Validate or format this JSON

One click loads this exact example into the tool — no copy-paste needed. Format it, validate it, explore the tree, or generate TypeScript types instantly.

Format Validate Explore tree Generate TypeScript Minify

Frequently Asked Questions

The model reads each tool's name and description (and parameter schema) and picks the best match for the user's request, then emits a tool call with arguments. Clear, specific descriptions are the single biggest factor in correct tool selection — treat them as prompts.

Mark them (e.g. a 'dangerous' flag), and require a human-in-the-loop confirmation or a permission check before executing. Never let an agent autonomously run irreversible actions like refunds or deletes without a guardrail.

Related JSON Examples

AI & LLMLLM Tool / Function CallFunction-calling JSON: a tool definition plus the model's tool_call with arguments.AI & LLMAI Agent Run TraceAgent run with goal, reasoning steps, tool calls, observations, and final output.AI & LLMMCP Message (Model Context Protocol)Model Context Protocol (MCP) JSON-RPC 2.0 message: tools/list and tools/call.