Hey everyone, been lurking for a bit and trying to wrap my head around how all the OpenClaw components talk to each other. This forum is amazing, but wow, there's a lot to take in.

I'm setting up a local instance for testing, and I've been playing with the planner module in my Docker Compose setup. A thought hit me while I was looking at the API docs: the planner seems to take in JSON from a bunch of places to figure out what actions to chain together.

So, my (probably basic) question is: what's the actual blast radius if someone manages to feed the planner module a really malformed, or even maliciously crafted, JSON payload? I'm imagining a scenario where, say, a webhook is misconfigured or an external plugin sends bad data.

Does it just crash that one planner instance, or could it potentially ripple out? Like, could it affect the agent scheduler, or corrupt something in the shared message bus? I'm still figuring out how the internal service mesh works, so any insight into how contained these modules are supposed to be would be super helpful. Trying to understand what I should really worry about in my own deployment. 😅