Engineering and IT teams that let AI coding agents work in repositories that hold credentials.
git clone https://github.com/JeongJaeSoon/agent-guard.gitThird party software. By copying, downloading or installing it you accept the Academy legal notice: you are responsible for reviewing its code and permissions before use.
Loading the file...
Agent Guard is a local guardrail that blocks risky .env reads, masks secret-like output and scans changed files for leaked secrets before they reach Git. It works with Claude Code, Codex, Git hooks, GitHub Actions and the shell, and uses gitleaks. It sends no telemetry and does not replace existing secret scanning.
Engineering and IT teams that let AI coding agents work in repositories that hold credentials. Agent Guard is a local guardrail that blocks risky .env reads, masks secret-like output and scans changed files for leaked secrets before they reach Git. It works with Claude Code, Codex, Git hooks, GitHub Actions and the shell, and uses gitleaks. It sends no telemetry and does not replace existing secret scanning.
Run this in your terminal: git clone https://github.com/JeongJaeSoon/agent-guard.git
Yes. The code is on GitHub (JeongJaeSoon/agent-guard) under the MIT licence.
It passed our automatic scan for credential theft, hidden instructions and risky install commands. Third party open source software. Sabemos AI does not maintain it. Check the code and permissions before you connect it to company data.