GitGuardian Just Gave AI Coding Agents Secret Detection Skills

GitGuardian

May 27, 2026

AI coding assistants like Claude Code and Cursor are helping developers write more code faster, but that also means more chances for secrets to slip into prompts, files, commits, and tool outputs. GitGuardian’s new open-source **agent-skills** repository teaches AI agents how to use **ggshield** directly inside the developer workflow: when to scan, how to read findings, and how to guide remediation for leaked credentials.

The repository includes a Claude Code plugin, Cursor rules and commands, and optional defense-in-depth with the ggshield agent hook, which can automatically scan prompts, tool calls, and tool outputs. It is public, MIT-licensed, and ready to demo, helping teams meet developers where they already work while reducing secret leakage in AI-assisted development.

Try GitGuardian AI Skills Now:
https://github.com/GitGuardian/agent-skills

00:00 Cold open

00:57 Introduction to skills

01:20 Which LLMs does this work with?

01:56 AI Skill builds on GitGuardian MCP

02:28 Use cases

03:41 MCP uses API vs AI Skills using ggshield

04:44 Do I need to reinstall ggshield?

05:24 Demo - Initial setup of the skill

06:52 Demo - Initial prompt "I have been breached..."

07:26 Demo - AI skill installing ggshield

09:19 Demo - Scanning for secrets and what it found