That's a valid point. The primary protection is file system isolation. The agent can't touch anything outside your project directory, and it won't have destructive access to most external APIs unless you provide keys. The threat model is really about preventing unintended file system changes. Full network sandboxing isn't really viable since Claude Code needs internet access to function.
I built this after realizing there's no structured way to improve workplace communication skills. Technical skills get tons of reps, but soft skills are seen as innate and expected to develop naturally.
Launched with interview practice and impromptu speaking scenarios that provide custom, structured feedback. The goal is deliberate practice for soft skills the same way we train technical skills.
I'm looking to add more modules, so I'd love to hear what soft skills challenges you've faced and what you'd want to work on.
reply