[Subtask] Develop error handler in Python

[debuggerone]

Subtask Overview

This subtask focuses on implementing a robust error handling mechanism in Python for the AgentM migration. The error handler should log errors to .var/log/error and manage access controls effectively.

Tasks

• [x] Develop an error handler in Python that logs errors to .var/log/error.
• [x] Implement access controls to ensure proper permissions for error logs.
• [x] Test the error handler to ensure it captures all relevant errors and logs them correctly.

Acceptance Criteria

• The error handler is implemented in Python and logs errors to .var/log/error.
• Access controls are in place to secure error logs.
• Error logging is tested and verified to be accurate and secure.

[Stevenic]

So developers are very bad about handling exceptions in JavaScript and services like OpenAI will throw some errors and return others. I elected to just return all errors which is why I have a standard AgentResponse class that has { completed, value, error } fields. I don't have an overly strong opinion about throwing versus returning errors other then you should just be consistent...

[debuggerone]

I'll make that configurable in install process

[debuggerone]

Errors are now simply output but also logged