Lack of error reporting / logs when agent communication fails

[diorwave]

When the bridge fails to connect to the underlying agent host or loses connection mid-session, there is little to no diagnostic feedback. This makes it difficult to debug connectivity problems or misbehaving tasks. Providing structured logs or error callbacks would improve developer visibility and troubleshooting.

[jcduro]

Thanks for raising this, this is a really valid point. Right now the bridge does very little in terms of surfacing connection or runtime failures to the user, which makes debugging pretty painful. A few concrete improvements that would help here:
​

Emit structured log events (with timestamps, connection state, and error codes) whenever the bridge fails to connect, loses the connection mid-session, or encounters protocol-level errors.
​

Expose an error/notification callback in the extension API so clients can subscribe to connection failures and display meaningful feedback in the UI (e.g. popups, status bar messages, or output channel entries).
​

Add an optional “debug mode” flag that increases verbosity and writes detailed traces to a dedicated output channel or log file for easier troubleshooting during development.
​

Happy to sketch a proposal (event schema + API surface) if you think this fits the project direction.