VibeKit
The safety layer for your coding agent
191 followers
The safety layer for your coding agent
191 followers
We built VibeKit to give developers control and visibility when running coding agents like Claude, Gemini, and OpenAI.
VibeKit runs your agent in an isolated Docker sandbox, redacts sensitive data before it leaves your machine, and gives you full observability into every action — file reads/writes, shell commands, API calls, outbound data.
Everything runs locally, offline, and it’s MIT-licensed open source.
This is the 2nd launch from VibeKit. View more
VibeKit CLI
Universal CLI wrapper that provides secure sandboxing, sensitive data redaction, and observability for Claude Code, Gemini CLI, or AI coding agent.
Free
Launch Team
VibeKit
what terminal u use in the video ?
@jay_emp0 this is Warp with its agents feature
Finden
Congrats on your launch! We launched Finden yesterday and just upvoted, would love your support.
Just tried wrapping a local Claude session with VibeKit, love that it redacts outbound data automatically and gives full transparency into what the agent's doing. Way less sketchy now
DeepTagger
Congratulations on the launch! I mean this is a real problem, but... like "REDACTED"? What will Claude Code think? You might want to communicate it BETTER to Claude Code that keys are being amended, OR, even better, replace those with DUMMY keys, and then REPLACE them back when it tries to output them.
AltPage.ai
Whoa, a safety layer for coding agents? That’s genius—my last project nearly went sideways from a rogue script. Super curious if VibeKit handles edge-case exceptions too!