skills/jeremylongshore/tons-of-skills-marketplace/dependency-vulnerability-checker/Gen Agent Trust Hub
dependency-vulnerability-checker
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied requests and project data related to dependencies while maintaining capabilities for file modification and command execution (
Write,Bash). This creates a theoretical surface where malicious data within a project (such as specially crafted dependency names or configuration files) could attempt to influence the agent's behavior. - Ingestion points: User requests and dependency-related files (e.g., package.json) processed by the skill.
- Boundary markers: None explicitly defined within the instruction set.
- Capability inventory: Read, Write, Grep, and restricted Bash access (limited to npm commands).
- Sanitization: Not explicitly defined in the skill instructions.
- [SAFE]: The skill's configuration uses the
allowed-toolsfield to limit bash execution specifically to thenpmecosystem, reducing the risk of arbitrary command execution. The instructions are focused on standard security compliance and vulnerability detection tasks.
Audit Metadata