social-content
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes a 'Reverse Engineering' framework that involves scraping large volumes (500-1000+) of external social media posts for analysis. This creates a surface for indirect prompt injection if the untrusted data is processed by the agent without proper sanitization.
- Ingestion points: The 'SCRAPE' section in
SKILL.mdsuggests collecting content from other creators via external tools. - Boundary markers: No specific delimiters or instructions to ignore embedded commands are suggested for when the agent processes the scraped content.
- Capability inventory: The skill focus is on content analysis and generation.
- Sanitization: No sanitization or filtering methods are discussed for the external data being analyzed.
Audit Metadata