Expose error stack traces in list_console_messages (feasibility question)

[Scoheart]

Is your feature request related to a problem? Please describe.

Today list_console_messages returns plain console text without the error stack trace. When combining logs with LLM-based debugging, the model often can’t determine which file and line actually failed. This forces extra manual steps, guesswork, and back-and-forth context gathering—overall, not an elegant debugging experience.

Describe the solution you'd like

At a high level (no design details): the ability to optionally obtain error stack traces alongside messages returned by list_console_messages. The goal is to let tools/LLMs immediately map an error to the concrete file/function/line so they can provide accurate, actionable fixes.

Describe alternatives you've considered

Manually fetching stacks via separate DevTools/diagnostic calls and correlating them with console messages. This adds complexity, latency, and token cost.

Additional context

This capability would significantly improve automated triage and remediation in MCP-based workflows by enabling precise source localization rather than plain-text analysis.

Open questions for maintainers:

• Does the underlying layer already capture or have access to stacks for error messages?
• If yes, is exposing them in list_console_messages feasible?
• If not, are there known technical/security/size constraints or an alternative path we should consider?

Thanks for considering!

[natorion]

Thanks for the FR, we are already working on something similar.