skill-builder
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill functions as a meta-instructional guide and template. It does not contain executable scripts, network requests, or commands that access sensitive files.\n- [INDIRECT_PROMPT_INJECTION]: The skill acts as a bridge for processing user-supplied data into structured agent instructions, which is a potential surface for indirect injection.\n
- Ingestion points: User input defining skill use cases, triggers, and steps (referenced throughout SKILL.md).\n
- Boundary markers: Not explicitly defined for the resulting generated output.\n
- Capability inventory: None; the skill itself does not utilize tools or perform automated actions.\n
- Sanitization: The skill provides proactive security advice by instructing the agent to omit XML angle brackets in metadata to prevent prompt injection.
Audit Metadata