Skip to content

agent-testing.md

Latest commit

 

History

History
84 lines (55 loc) · 2.54 KB

agent-testing.md

File metadata and controls

84 lines (55 loc) · 2.54 KB

Agent Testing Feature

Test your AI agents with real-time streaming conversations.

Overview

The agent testing feature allows you to interact with your configured agents in a chat interface before deploying them to production tasks. This helps validate agent behavior, system prompts, and model configurations.

How to Test an Agent

From Agent Card

  1. Navigate to the Agents page
  2. Find the agent you want to test
  3. Click the Test Agent button at the bottom of the agent card
  4. Or click the three-dot menu and select Test Agent

From Agent Details

  1. Click on an agent to view its details
  2. Click the Test Agent button in the header
  3. The test modal will open

Test Interface

The test interface provides:

  • Real-time streaming responses: See the agent's response as it's generated
  • Chat history: View the conversation history during the test session
  • Message input: Send messages to test different scenarios
  • Error handling: Clear error messages if something goes wrong

Keyboard Shortcuts

  • Enter: Send message
  • Shift + Enter: New line in message

What Gets Tested

When you test an agent, the system uses:

  • Agent's configured system prompt
  • Selected LLM model and provider
  • Temperature, max tokens, and other model parameters
  • Agent's capabilities and skills (if configured)

Tips for Effective Testing

  1. Test edge cases: Try unusual inputs to see how the agent handles them
  2. Verify system prompt: Check if the agent follows its configured instructions
  3. Test different scenarios: Use various types of questions relevant to the agent's role
  4. Check response quality: Evaluate if the model and temperature settings are appropriate

Limitations

  • Test sessions are temporary and not saved
  • Test conversations don't affect agent statistics or task counts
  • Memory and knowledge base access may be limited in test mode

Troubleshooting

"Failed to load available providers"

  • Check that at least one LLM provider is configured in Settings
  • Verify the provider is running and accessible
  • Check network connectivity

"No response from agent"

  • Verify the selected model is available
  • Check LLM provider logs for errors
  • Ensure the model has sufficient resources

Streaming stops unexpectedly

  • Check network connection
  • Verify max tokens setting isn't too low
  • Review backend logs for errors

Related Documentation