shifting-perspective
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill transmits user-provided content, including URLs and YouTube video identifiers, to the InfraNodus MCP server (
mcp.infranodus.com) for processing. This is the primary function of the skill and utilizes the vendor's own infrastructure. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, creating a surface for potential indirect prompt injection attacks.
- Ingestion points: Data from URLs and transcripts is ingested by the
optimize_text_structuretool described inSKILL.md. - Boundary markers: Absent; the skill does not explicitly isolate the ingested content with delimiters.
- Capability inventory: The skill uses tool outputs to generate research questions, identify content gaps, and suggest perspective shifts.
- Sanitization: Absent; the agent interprets findings from the tool without specific content filtering instructions.
Audit Metadata