Agent troubleshooting

Use this guide to diagnose and fix common Agent issues.

Quick triage

  1. Confirm the model provider is enabled in Agents Config
  2. Confirm the Agent is connected to a valid model
  3. Confirm required tools are enabled
  4. Run a simple known-good test prompt

Common problems

ProblemLikely causeFix
Agent does not respondProvider/model is not configured correctlyRe-check API credentials, model name, and provider status
Agent fabricates valuesMissing tool-use rule or no read tool enabledAdd strict tool rules and enable required read tools
Agent updates data unexpectedlyWrite rules are too vagueAdd explicit write constraints and refusal rules
Automation parsing failsOutput format changes between responsesEnforce a stable JSON schema in instructions
Self-hosted AI requests failNetwork egress restrictionsFollow Self-hosted AI features requirements

Debugging checklist

When behaviour is inconsistent:

  • Reduce prompt complexity and retest
  • Verify the tool list in the Agent context
  • Add explicit examples of valid and invalid behaviour
  • Test each instruction change against a fixed regression set

When to change instructions vs tools

Change instructions when:

  • Behaviour is mostly correct but inconsistent
  • Refusal or escalation logic needs tightening

Change tools when:

  • The Agent cannot access required data
  • The Agent has risky capabilities it should not have

Related guides