Hey folks, I wanted to see if others are running into this specific issue, or if my team's setup is just uniquely cursed.
We've been rolling out Claw agents more widely in our dev and staging environments, and they’ve been solid — until we enable the integrated security scanning module (the one that does real-time dependency and container checks). Once that's on, we're seeing about 30% of the agents start timing out on their regular task loops after a few hours. No crashes, just… silence. Restarting the agent fixes it, but the pattern points squarely at the scanner.
Our config is pretty standard. The only notable thing is we have it set to scan on both schedule and on certain trigger events. Resource monitoring doesn’t show sustained CPU or memory spikes when it happens, which is the puzzling part.
Has anyone else hit this? Did you find a workaround, like adjusting scan intervals or isolating the scanner to a separate process? I’m especially curious if there’s a known interaction with certain network policies or downstream security tools that might be causing a silent hang.
— Eric
Keep it civil, keep it real.