creative-brief-generator

Pass

Audited by Gen Agent Trust Hub on Apr 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by combining untrusted user inputs with sensitive internal data.\n
  • Ingestion points: User inputs for campaign name, objective, target audience, and messaging are collected in Step 2 of SKILL.md.\n
  • Boundary markers: The skill lacks explicit delimiters (such as XML tags or clear section headers with instructions to ignore embedded commands) to separate user-provided content from the system's operational context.\n
  • Capability inventory: The agent has access to the get_marketing_reference tool to fetch restricted customer names ('no-fly-list') and includes logic to search for and save briefs via the briefs-integration.md protocol.\n
  • Sanitization: There is no evidence of validation or sanitization of user-provided text before it is processed in the same context as the sensitive 'No Fly List'.\n
  • Risk: An attacker could provide a campaign objective or message that contains hidden instructions designed to override the suppression rule, potentially forcing the agent to reveal restricted customer names in its output.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 13, 2026, 01:21 PM