
HOL Guard
Review risky AI coding agent actions before they run
Launched 10/17/2026
2 views
0 comments
About HOL Guard
HOL Guard is a local-first runtime checkpoint for AI coding agents such as Claude Code, Codex and Cursor. It reviews risky actions (shell commands, MCP tool calls, secret reads, package installs) before they run; your local policy allows, blocks or asks you first, and decisions stay in a local log. Not a network gateway or firewall, not a complete prompt-injection defense; scans are not a safety guarantee. Apache-2.0 core, optional paid Guard Cloud. Freemium. Made by HOL. https://hol.org/guard