Skip to content

Troubleshooting

Start here when something is not working. Follow the steps in order.

First Steps

1. Check if the daemon is running

bash
triggerfish status

If the daemon is not running, start it:

bash
triggerfish start

2. Check the logs

bash
triggerfish logs

This tails the log file in real time. Use a level filter to cut through the noise:

bash
triggerfish logs --level ERROR
triggerfish logs --level WARN

3. Run diagnostics

bash
triggerfish patrol

Patrol checks whether the gateway is reachable, the LLM provider responds, channels are connected, policy rules are loaded, and skills are discovered. Any check marked CRITICAL or WARNING tells you where to focus.

4. Validate your config

bash
triggerfish config validate

This parses triggerfish.yaml, checks required fields, validates classification levels, and resolves secret references.

Troubleshooting by Area

If the first steps above did not point you to the problem, pick the area that matches your symptoms:

  • Installation - install script failures, build-from-source issues, platform problems
  • Daemon - service will not start, port conflicts, "already running" errors
  • Configuration - YAML parse errors, missing fields, secret resolution failures
  • Channels - bot not responding, auth failures, message delivery issues
  • LLM Providers - API errors, model not found, streaming failures
  • Integrations - Google OAuth, GitHub PAT, Notion API, CalDAV, MCP servers
  • Browser Automation - Chrome not found, launch failures, navigation blocked
  • Security & Classification - write-down blocks, taint issues, SSRF, policy denials
  • Secrets & Credentials - keychain errors, encrypted file store, permission problems

Still Stuck?

If none of the guides above resolved your issue:

  1. Collect a log bundle
  2. Read the filing issues guide
  3. Open an issue on GitHub