competitor-ad-teardown
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or data exfiltration attempts were detected. The skill uses natural language to guide the agent through a manual research process using a well-known service (Meta Ad Library).
- [NO_CODE]: The skill contains no executable scripts (Python, Node.js, Shell, etc.), which eliminates the risk of remote code execution or unauthorized system modifications.
- [PROMPT_INJECTION]: The skill involves processing external data from the Meta Ad Library (facebook.com/ads/library).
- Ingestion points: Publicly available ad copy and creative descriptions from competitors.
- Boundary markers: The prompt provides a structured extraction framework (Who, Pain, Promise, Register) which helps maintain task focus.
- Capability inventory: No subprocess, file-write, or non-whitelisted network capabilities are used; it relies purely on the agent's native browsing ability.
- Sanitization: The instructions explicitly state 'Never copy their copy; the output is directions, not text', which serves as a defensive instruction against indirect injection from ad content.
Audit Metadata