do-distill-to-skill
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill contains guidelines and templates for information processing. It does not attempt to bypass safety filters, exfiltrate credentials, or establish persistence. All instructions are focused on the legitimate task of knowledge distillation.
- [INDIRECT_PROMPT_INJECTION]: The skill's primary workflow involves processing untrusted external data, which is an inherent surface for indirect prompt injection attacks.
- Ingestion points: External URLs, blog posts, articles, and GitHub repositories (SKILL.md, Workflow Step 1).
- Boundary markers: The instructions suggest a structured output template but do not explicitly mandate the use of delimiters or 'ignore' instructions for the raw input data.
- Capability inventory: The skill facilitates the reading and analysis of external content, often used in conjunction with file-writing capabilities (mentioned via the
skill-creatorcompanion skill). - Sanitization: The 'Distillation' process itself acts as a cognitive filter, as it instructs the agent to extract universal principles and discard conversational or promotional prose, which inherently reduces the impact of embedded instructions.
Audit Metadata