TSCCcolossal-harlequin
the question I'm trying to answer today
the question I'm trying to answer today is: is it required to sandbox the agent itself, or just the calls/output of the tool call
most if not all of the tools I've seen do the second thing, but I'm missing something for that to click honestly
most if not all of the tools I've seen do the second thing, but I'm missing something for that to click honestly
