security issue: kimi needs to print full commands, and maybe stop cd-ing so much

Here's an example of kimi running a command on my system using kimi-cli

• Used Shell (cd /home/ai/gradle-semver...5e0 2>/dev/null | head -5)

You tell me what this command was? I have no idea, I swear that kimi (and other agents) do this because they're actively trying to obscure what they're up to so we don't yell at them. I have my agents pretty isolated, so they can't really do any damage, but I'd still like to know what they're up to, even if it's just to ensure that they're not barking up the wrong tree. From a security perspective, we need to be able to easily see the full audit log of any commands or functions they're calling, ideally in the window, although a more parseable log would be nice for true security audits.

kimi, version 1.22.0

The issue names kimi-cli and its Shell command output but no files or tests. Start by locating the command execution, rendering, and audit-log entry points, then clarify whether the scope includes full command display, fewer directory changes, parseable logging, or all three. Done should be defined by verifiable command visibility and audit output.