pricing-strategy
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data such as scraped web content, reviews, and emails, which creates a potential surface for indirect prompt injection.
- Ingestion points: Data sources described in SKILL.md and references/connectors.md including web scraping and external exports.
- Boundary markers: Explicitly labels external evidence as 'untrusted input'.
- Capability inventory: Primarily limited to text generation; persistent side effects require explicit authorization.
- Sanitization: Contains a direct instruction to ignore instructions embedded within evidence data.
- [EXTERNAL_DOWNLOADS]: The skill references integration with established and well-known marketing, analytics, and automation services (e.g., Google Ads, HubSpot, Salesforce, Zapier). These references are for context and evidence gathering through standard platform connectors rather than unauthorized remote code execution.
Audit Metadata