ad-brief
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing untrusted data from external and local sources.
- Ingestion points: The skill ingests untrusted data from external URLs during its research phase, project-level metadata (e.g., package.json, README.md), and user-provided reference materials like brand guidelines.
- Boundary markers: The instructions do not include the use of delimiters or specific directives for the agent to ignore instructions that might be embedded in the researched content.
- Capability inventory: The skill utilizes the agent's ability to read project files and write the resulting ad brief to the file system at .agents/ad-brief.md.
- Sanitization: No sanitization or validation mechanisms are specified for the external content before it is incorporated into the brief.
Audit Metadata