shared-security-auth-security
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes an example script ('examples/dependency-security.md') intended for CI/CD pipelines that executes the 'bun audit' command using 'child_process.exec'. This is a routine operation for automated vulnerability scanning within development workflows.
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to provide guidance on processing untrusted data safely, creating a potential attack surface if user-provided content is handled without follow-up sanitization. 1. Ingestion points: The skill instructions and examples assume the agent will interact with and review user-provided source code and security configurations as specified in 'SKILL.md'. 2. Boundary markers: The skill utilizes structured markdown tags such as '<critical_requirements>' and '' to isolate core instructions from descriptive content. 3. Capability inventory: Code examples within the skill demonstrate capabilities such as making network requests via 'fetch' ('examples/core.md') and executing system commands for auditing ('examples/dependency-security.md'). 4. Sanitization: The skill explicitly requires and provides implementation details for the sanitization of all user input using 'DOMPurify' before any HTML injection occurs, mitigating XSS and related injection risks.
Audit Metadata