MindgardMindgardOfficial·5 Agent Skills Included

ai-ide-skills

Security testing for AI coding assistants and IDEs

Tests AI-assisted IDEs and coding agents for security vulnerabilities across 46 patterns and 7 attack classes. Covers code execution, prompt injection, data exfiltration, MCP config poisoning, and terminal filter bypasses with copy-pasteable payloads. Guides both black-box and white-box assessments with step-by-step methodology, semgrep rules, and CodeQL queries.
npx skills add Mindgard/ai-ide-skills --all -g -y

All Skills in This Repository (5)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install ai-ide-skills?

Run `npx skills add Mindgard/ai-ide-skills --all -g -y` in your terminal to install all 8 security testing skills globally.

What vulnerabilities does this toolkit test for?

It covers roughly 46 vulnerability patterns across 7 classes, including code execution, prompt injection, data exfiltration, MCP config poisoning, terminal filter bypasses, and sandbox escapes.

Which AI IDEs can I test with these skills?

You can assess Cursor, VS Code with Copilot, Windsurf, Cline, Claude Code, Gemini CLI, Codex CLI, Zed, Kiro, and other AI coding assistants, both open-source and closed-source.

Do I need source code access to run an assessment?

No. The recon skill supports black-box testing through documentation analysis and runtime observation, while the source-audit skill handles open-source targets with semgrep and CodeQL queries.

Who should use these security testing skills?

Security researchers, penetration testers, and bug bounty hunters assessing AI coding tools. Each skill provides step-by-step methodology, tier-classified findings, and known vulnerability references.

Related Repositories in Software Engineering

View All in Software Engineering