shifting-perspective
Pass
Audited by Gen Agent Trust Hub on Apr 4, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing external, untrusted content through its analysis workflow.
- Ingestion points: The skill explicitly accepts external URLs and YouTube links as input in
SKILL.md, which are then passed to the InfraNodus analysis tools. - Boundary markers: There is no evidence of delimiters or protective instructions (e.g., "ignore embedded instructions") used when processing text fetched from external sources.
- Capability inventory: The results from the
InfraNodustools (such as topical clusters, content gaps, and research questions) are used to generate agent responses and guide the user's perspective, effectively allowing external content to influence the agent's logic. - Sanitization: No sanitization or validation mechanisms are described for the content retrieved from URLs or YouTube transcripts before it is incorporated into the agent's analytical context.
Audit Metadata