[Feature]: Agentic feedback loop

[cjlryan]

Feature Description

Instead of just streaming logging, a "capture_log_event" tool that waits for a specified phrase or statement to appear in the log, then returns the surrounding logs to the agent to form an agentic feedback loop would be super useful.

Use Cases

• Building any app

Example Interactions

You: Make [some feature], use logging to check it works.

AI: Making feature...

func someFeature() {
     print("Something broke...")
}

AI: Checking logs...

Tool Call:

capture_log_event(process: "SomeProcess", phrase: "Something broke...", max_logs_to_include: 10)

Tool Result:

Everything's fine
Something bad happens
Something broke...

AI: Hmm, that didn't work, trying [something else]

[github-actions]

This issue has been inactive for 30 days. It will be closed in 7 days if no further activity occurs. Add a comment to keep it open, or apply the no-stale label.

[cameroncooke]

That's an interesting idea, my concern is that the agent will have to be interactively running the log tool until the tool finds a match or times out, this is somewhat alleviated by background bash support some agents have though not all support this. I'm not closed off to this idea though just needs some thought. If you want to take a stab at it and open a PR that would be welcome!