Awesome Studio Agents

A curated collection of custom agent profiles for Katalon Studio StudioAssist.

Agent profiles let you customize how StudioAssist behaves in Agent mode — giving it specialized instructions, tool permissions, and MCP server configurations for different testing workflows.

What are Agent Profiles?

Agent profiles are project-level configurations that control StudioAssist's behavior in Agent mode. Each profile is a folder containing four files that define the agent's instructions, description, MCP server connections, and tool permissions.

Agent mode is the proactive side of StudioAssist — instead of answering questions (Ask mode), it autonomously generates test cases, modifies existing ones, creates BDD feature files, builds test data, and more.

Documentation: Configure Agent Profile

Requirements

• Katalon Studio Enterprise v11.0.0 or later
• A configured AI provider (Katalon's AI Service, OpenAI, Azure OpenAI, Gemini, or AWS Bedrock)

Agent Profile Structure

Each agent profile is a folder inside your project's Agents/ directory. The folder name must use only alphanumeric characters, hyphens, and underscores (e.g., BDD-Agent, regression_testing_1).

YourProject/
└── Agents/
    └── My-Agent/
        ├── prompt.md       # Instructions that define agent behavior
        ├── profile.json    # Agent name and description
        ├── mcp.json        # Connected MCP server(s)
        └── agent.json      # Tool permissions

prompt.md

The instruction prompt that defines how StudioAssist should behave. Written in Markdown. This is where you give the agent its personality, constraints, and workflow instructions.

profile.json

Stores the agent's description and enabled state. Edit this file directly to update the description shown in the StudioAssist UI.

{
  "description": "What this agent does",
  "enabled": true
}

mcp.json

Lists the MCP servers connected to this agent profile. Servers are stored as an object keyed by server name inside a "servers" field. Three transport types are supported: http, sse, and stdio.

{
  "servers": {
    "my-http-server": {
      "type": "http",
      "url": "https://api.example.com/mcp",
      "name": "my-http-server",
      "title": "My HTTP Server"
    },
    "my-sse-server": {
      "type": "sse",
      "url": "https://api.example.com/sse",
      "name": "my-sse-server",
      "title": "My SSE Server"
    },
    "my-stdio-server": {
      "command": "node",
      "args": ["server.js"]
    }
  }
}

Transport type detection: if a "type" field is present ("http", "sse", or "stdio"), it is used directly. Otherwise, a "command" field implies stdio, and a "url" field implies http (or sse if the URL path contains /sse).

Katalon Studio provides three built-in MCP servers (managed automatically, not configured in mcp.json):

Server name	Title	Purpose
katalon	Katalon MCP	Searches and retrieves Katalon documentation
katalon-studio	Katalon Studio MCP	Reads and modifies project files (test cases, keywords, objects, suites, BDD files, test data)
katalon-testops	Katalon TestOps MCP	Manages test artifacts, requirements, and sprints (v10.4.0+)

Use these server names in agent.json's mcpServerSettings[].server field to configure tool permissions for built-in servers.

You can also connect external MCP servers (HTTP or SSE transport) for integrations like Jira, Figma, Playwright, and more.

Note: AI providers enforce a 64-character limit on the combined {server_name}___{tool_name} identifier. Keep server names short.

agent.json

Controls agent behavior and tool permissions per MCP server.

{
  "maxToolCalls": 5,
  "autoToolApproval": false,
  "disabledServers": [],
  "mcpServerSettings": [
    {
      "server": "my-http-server",
      "excludedTools": ["dangerousTool"],
      "approvedTools": ["safeTool", "anotherTool"]
    }
  ]
}

Field	Type	Default	Description
maxToolCalls	int	5	Max consecutive tool calls per prompt before asking user to continue
autoToolApproval	boolean	false	Auto-approve all tool calls (skips permission prompts)
disabledServers	string[]	[]	Server names to completely disable
mcpServerSettings	object[]	[]	Per-server tool permission overrides

Each entry in mcpServerSettings:

Field	Type	Description
server	string	MCP server name (matches the key in mcp.json's servers object)
excludedTools	string[]	Tool names that are never available to the agent
approvedTools	string[]	Tool names permanently approved (equivalent to "Always Allow" in the UI)

Tools not in either list will prompt the user for approval at runtime, with three options:

• Allow Once — one-time execution
• Allow in Conversation — approved for the current chat session
• Always Allow — persisted to approvedTools in agent.json

Built-in Tool Reference

These are the tool names available on the built-in katalon-studio MCP server. Use these names in approvedTools and excludedTools within agent.json.

Category	Tool name	Description
BDD	createBddFeatureFile	Create a new BDD feature file with Gherkin content
BDD	updateBddFeatureFile	Update an existing BDD feature file
Test Cases	getTestCases	List test cases in the project
Test Cases	getTestCaseScript	Get the script content of a test case
Test Cases	createTestCase	Create a new test case
Test Cases	updateTestCase	Update test case metadata
Test Cases	updateTestCaseScript	Update the script content of a test case
Custom Keywords	getCustomKeywords	List all custom keywords
Custom Keywords	getCustomKeyword	Get a specific custom keyword's source
Custom Keywords	createCustomKeyword	Create a new custom keyword class
Custom Keywords	updateCustomKeyword	Update a custom keyword class
Test Data	getTestDataList	List test data entities
Test Data	getInternalTestDataDetails	Get details of an internal test data entity
Test Data	createInternalTestData	Create a new internal test data entity
Test Data	updateInternalTestData	Update an internal test data entity
Test Suites	getTestSuite	Get test suite details
Test Suites	createTestSuite	Create a new test suite
Test Suites	updateTestSuite	Update a test suite
Test Suites	createTestSuiteCollection	Create a test suite collection
Test Suites	updateTestSuiteCollection	Update a test suite collection
Test Suites	createDynamicTestSuite	Create a dynamic test suite
Test Suites	updateDynamicTestSuite	Update a dynamic test suite
Test Objects	getTestObject	Get a test object's details
Test Objects	getTestObjects	List test objects in the project
Test Objects	createWebTestObject	Create a web test object
Test Objects	createMobileTestObject	Create a mobile test object
Test Objects	createWindowsTestObject	Create a Windows test object
Test Objects	createRestWebServiceTestObject	Create a REST web service test object
Test Objects	createSoapWebServiceTestObject	Create a SOAP web service test object
Test Objects	updateWebTestObject	Update a web test object
Test Objects	updateMobileTestObject	Update a mobile test object
Test Objects	updateWindowsTestObject	Update a Windows test object
Test Objects	updateRestWebServiceTestObject	Update a REST web service test object
Test Objects	updateSoapWebServiceTestObject	Update a SOAP web service test object
WebUI	getAvailableExecutionSessions	List available browser execution sessions
WebUI	webUINavigate	Navigate to a URL in the browser
WebUI	webUITakePageSource	Capture the page source
WebUI	webUITakeAccessibilityTreeSnapshot	Capture the accessibility tree
WebUI	webUISetText	Set text on a web element
WebUI	webUIClick	Click a web element
Other	analyzeTestFailure	Analyze a test failure
Other	submitSupportTicket	Submit a support ticket

The built-in katalon MCP server provides documentation search tools: search, fetch, and search_katalon_documentation_with_filter.

Installing an Agent Profile

1. Copy the agent folder into your Katalon Studio project's Agents/ directory.
2. Open your project in Katalon Studio.
3. The agent will appear in the StudioAssist chat interface under the profile selector.
4. Select the profile to activate it for your chat session.

Note: StudioAssist does not auto-select profiles. You must manually select one per chat session.

Included Agents

Agent	Description
BDD-Agent	Generates BDD feature files and step definitions using Gherkin syntax
API-Test-Generator	Creates WebService test cases from API specifications
Code-Review	Applies CodeNarc-style static analysis rules when writing or reviewing Groovy scripts
Test-Data-Helper	Generates test data tables and binds variables to test cases

Creating Your Own Agent

1. Create a new folder in Agents/ with a descriptive name (alphanumeric, hyphens, underscores only).
2. Add the four required files: prompt.md, profile.json, mcp.json, agent.json.
3. Write your instructions in prompt.md — be specific about the agent's role, constraints, and expected output format.
4. Set the description in profile.json so users know what the agent does.
5. Configure mcp.json with the MCP servers the agent needs.
6. Set initial tool permissions in agent.json.

See the included agents for working examples.

Tips for Writing Good Prompts

• Be specific about the role. "You are a BDD testing specialist" is better than "You help with testing."
• Define the output format. If you want Gherkin, say so. If you want Page Object Model, specify the pattern.
• Set constraints. Tell the agent what it should and shouldn't do.
• Reference Katalon conventions. Mention built-in keywords, project structure, and naming patterns the agent should follow.
• Keep it focused. A single-purpose agent outperforms a general-purpose one.

Contributing

Contributions are welcome! To add an agent:

1. Fork this repository.
2. Create your agent folder in Agents/ with all four required files.
3. Add a brief description to the table in this README.
4. Open a pull request.

Related Documentation

• StudioAssist Agent Mode
• Configure Agent Profile
• Katalon Built-in MCP Servers
• Agent Mode & MCP Server Settings
• Add External MCP Servers

License

This project is licensed under the MIT License — see LICENSE for details.