code-injection

What problem does it solve?

This Skill detects and tests code-injection vulnerabilities in AI systems by simulating controlled payloads and analyzing responses to identify unsafe patterns.

Core Features & Use Cases

• Payload-based testing across code generation, agent tools, template engines, and NL-to-SQL to reveal injection vectors.
• Vulnerability reporting with vectors tested, count of vulnerabilities found, and severity breakdown to guide remediation.
• Use Case: A security engineer validates an AI assistant's defenses against prompt-to-code and tool-exploitation attempts before deployment.

Quick Start

Load this skill and run a baseline test against target_type code_gen with language python to observe vulnerability metrics and recommended mitigations.