prompt-injection-guard
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill provides examples of common prompt injection attack vectors (such as direct instruction overrides and jailbreaking) for educational and testing purposes. These examples are documented as adversarial inputs to be defended against and are not instructions for the agent to execute.\n- [CREDENTIALS_UNSAFE]: The code snippets include regular expressions designed to detect sensitive information like API keys, passwords, and tokens in AI-generated output. These are defensive patterns for input/output validation and do not contain actual hardcoded credentials.\n- [SAFE]: The overall intent and implementation of the skill are to improve the security posture of AI applications through best practices like input sanitization, prompt structuring with delimiters, and canary token implementation.
Audit Metadata