Skip to content

Troubleshooting

This guide helps resolve common issues with the Agent Starter Pack.

Authentication Issues

For detailed information on authentication with Vertex AI, visit the official documentation.

"Could not find credentials" or "Could not find project" Error

Problem: Missing credentials error with Vertex AI.

Solution:

  1. Log in to Google Cloud: gcloud auth login --update-adc
  2. Set the correct project:
    bash
    gcloud config set project YOUR_PROJECT_ID
    gcloud auth application-default set-quota-project YOUR_PROJECT_ID

Vertex AI API Not Enabled

Problem: Operations fail because the Vertex AI API is not enabled in your project.

Solution:

  1. Enable the Vertex AI API:

    bash
    gcloud services enable aiplatform.googleapis.com
  2. Verify the API is enabled:

    bash
    gcloud services list --filter=aiplatform.googleapis.com

Permission Denied Errors

Problem: "Permission denied" errors with Google Cloud APIs.

Solution: Ensure your user or service account has the necessary IAM roles. For example, for Vertex AI, you often need roles/aiplatform.user. Grant roles using the gcloud projects add-iam-policy-binding command or the Cloud Console.

Command Not Found: agent-starter-pack

Problem: "Command not found" error after installation.

Solution:

  1. Verify installation:
    bash
    pip list | grep agent-starter-pack
  2. Check PATH:
    bash
    echo $PATH
  3. Reinstall if needed:
    bash
    pip install --user agent-starter-pack
  4. For pipx:
    bash
    pipx ensurepath
    source ~/.bashrc  # or ~/.zshrc

Project Creation Issues

Project Creation Fails

Problem: agent-starter-pack create fails.

Solution:

  1. Check Error Messages: Examine output for clues.
  2. Write Permissions: Ensure write access to the directory.
  3. Project Name: Use lowercase letters, numbers and hyphens only.
  4. Debug Mode: Consider using debug mode to get more detailed error information:
    bash
    agent-starter-pack create my-project-name --debug

Issues with Agent Engine

Consider leveraging the public product documentation

Getting More Help

If issues persist:

  1. Check GitHub Issues: Search for existing Github issues in the agent-starter-pack Github repository.

  2. File a New Issue: Provide:

    • Problem description.
    • Steps to reproduce.
    • Error messages (preferably run with --debug flag for detailed logs).
    • Environment: OS, Python version, agent-starter-pack version, installation method, shell.

Released under the Apache 2.0 License.