NYC

make-skill-template

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • Prompt Injection (MEDIUM): The skill possesses an indirect prompt injection surface (Category 8). It ingests untrusted user input to generate executable-like markdown files and directory structures.
  • Ingestion points: User prompts triggered by phrases like "create a skill" or "make a new skill" serve as inputs for the generated content.
  • Boundary markers: The template lacks explicit boundary markers or instructions to the agent to ignore potentially malicious instructions embedded within the user's request for a new skill.
  • Capability inventory: The skill allows the agent to create directories and write content to SKILL.md files, which define the behavior of future agent interactions.
  • Sanitization: There is no evidence of sanitization or validation of the input strings before they are interpolated into the generated skill's frontmatter or body.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 06:13 AM