HOL Guard
Local checks on AI coding-agent shell, secrets, MCP calls, and installs before they run
About
HOL Guard sits on your machine and reviews what an AI coding agent is about to do: run a shell command, read a credential, call an MCP tool, or install a package. You set the policy so each action is allowed, blocked, or held for your approval. The core is open source under Apache-2.0. Install with pipx install hol-guard && hol-guard init. Signed Windows desktop installer available. Optional paid Guard Cloud for teams. Not a gateway or network firewall; scans are not a safety guarantee or complete prompt-injection preventer. Made by HOL.
Why try HOL Guard?
The problem: AI coding agents can run risky shell commands, read secrets, call MCP tools, or install packages before you notice
Built for: Developers and teams using AI coding agents who want local policy control on their machine
0 comments
No comments yet — be the first to react.