marketing-psychology
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions define workflows for ingesting external data through scraping and enrichment via SandBase API tools, creating a potential surface for indirect prompt injection.
- Ingestion points:
SKILL.mddescribes using SandBase for "current external evidence, search, scraping, [and] enrichment." - Boundary markers: While the skill advises separating facts from interpretation, it does not mandate specific delimiters (e.g., XML tags or backticks) for untrusted data.
- Capability inventory: The skill includes network access via
sandbase_runand file inspection capabilities. - Sanitization: No specific sanitization or input validation logic is prescribed for handling the retrieved external content.
Audit Metadata