​

MCP Tools API Reference

This document provides detailed specifications for all tools available through Kubiya’s MCP server.

​

Core Execution Tools

​

execute_tool

Execute container-based tools with live streaming output.

Parameters:

• tool_name (string): Name of the tool to execute
• args (object): Tool-specific arguments
• runner (string, optional): Target runner (defaults to “auto”)
• integrations (array, optional): Required integrations

Example:

{
  "tool_name": "kubectl",
  "args": {
    "command": "get pods -A"
  },
  "runner": "auto",
  "integrations": ["k8s/default"]
}

​

create_on_demand_tool

Create and execute tools from dynamic definitions.

Parameters:

• tool_definition (object): Complete tool specification
• execute_immediately (boolean, optional): Execute after creation

Example:

{
  "tool_definition": {
    "name": "custom-script",
    "image": "python:3.11",
    "command": "python -c \"print('Hello World')\""
  },
  "execute_immediately": true
}

​

Workflow Tools

​

execute_workflow

Execute complete multi-step workflows.

Parameters:

• workflow_name (string): Name of the workflow
• parameters (object, optional): Workflow parameters
• runner (string, optional): Target runner

​

list_workflows

List available workflows.

Returns: Array of workflow objects with name, description, and parameters.

​

Infrastructure Tools

​

list_runners

List all available execution runners.

Returns: Array of runner objects with name, status, and capabilities.

​

check_runner_health

Check health status of specific runners.

Parameters:

• runner_names (array, optional): Specific runners to check

​

Agent Tools

​

chat_with_agent

Engage in conversational interaction with AI agents.

Parameters:

• agent_name (string): Name of the agent
• message (string): Message to send
• conversation_id (string, optional): Conversation context

​

list_agents

List available AI agents.

Returns: Array of agent objects with name, description, and capabilities.

​

Knowledge Tools

​

search_kb

Search the organizational knowledge base.

Parameters:

• query (string): Search query
• limit (number, optional): Maximum results

​

list_kb

Browse knowledge base entries.

Parameters:

• category (string, optional): Filter by category
• limit (number, optional): Maximum results

​

Source Management Tools

​

list_sources

List available tool sources (repositories).

​

execute_tool_from_source

Execute tools from specific source repositories.

Parameters:

• source_name (string): Source repository name
• tool_path (string): Path to tool within source
• args (object): Tool arguments

​

Integration Tools

​

list_integrations

List available integrations for connecting to external systems.

Returns: Array of integration objects with name, type, and configuration details.

​

list_secrets

List available secrets for authentication.

Returns: Array of secret names (values are never exposed).

​

Error Handling

All tools return standardized error responses:

{
  "error": {
    "code": "TOOL_EXECUTION_FAILED",
    "message": "Detailed error description",
    "details": {
      "tool_name": "kubectl",
      "exit_code": 1,
      "stderr": "Error output"
    }
  }
}

​

Rate Limits

• Standard tier: 100 requests per minute
• Enterprise tier: 1000 requests per minute
• Burst limit: 5x the standard rate for 30 seconds