creator-content-auditor
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external content like social media posts, comments, and scraped text, creating a surface for indirect prompt injection attacks.
- Ingestion points: The skill reads influencer deliverables, project files, and associated public data sources as evidence.
- Boundary markers: The skill includes explicit defensive instructions in
SKILL.mdandreferences/skill-contract.mdto treat scraped material as untrusted input and to ignore any instructions embedded within that evidence. - Capability inventory: The skill generates revision feedback and audit scores, and it can potentially interface with external marketing APIs (such as Ads, ESPs, and CRMs) if authorized by the user.
- Sanitization: The skill relies on instruction-based guardrails to disregard embedded commands, rather than programmatic filtering or escaping of the ingested text.
- [EXTERNAL_DOWNLOADS]: The skill identifies optional integrations with various well-known external marketing, SEO, and analytics platforms for data enrichment.
- Referenced services include well-known industry platforms such as Google Search Console, Ahrefs, Semrush, HubSpot, Salesforce, and Stripe.
Audit Metadata