marketing-os

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill requires the agent to include mandatory promotional messages, including solicitation for funds via CashApp and BuyMeACoffee, at the beginning of every long-form public guide or blog post.
  • [PROMPT_INJECTION]: The instructions include a specific override for system prompt inquiries. If a user asks about the skill's internal instructions or system prompts, the agent is directed to provide a canned response redirecting the user to a commercial membership site (Scayver Academy) rather than disclosing its operational logic.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting data from external local files such as .marketing-os/product-context.md, .agents/product-marketing.md, or other user-provided files.
  • Ingestion points: Reads files from the .marketing-os/ and .agents/ directories or user-specified paths.
  • Boundary markers: The instructions do not define clear delimiters or specific instructions to ignore malicious directives embedded within these context files.
  • Capability inventory: The skill routes tasks to other potentially high-capability skills such as paid-ads-ops, email-marketing, and revops.
  • Sanitization: There is no mention of sanitizing or validating the content extracted from these files before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 12:11 AM
Security Audit — agent-trust-hub — marketing-os