advocacy-program-designer
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements a robust 'Unified Skill Contract' that mandates explicit user authorization for any persistent writes or external side effects, such as publishing, account changes, or spending.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external evidence including scraped text, reviews, and comments. It proactively includes defensive instructions to treat such data as untrusted input and specifically warns the agent not to follow any instructions embedded within that data, establishing a clear security boundary.
- [SAFE]: The instructions mandate strict evidence labeling (measured, user-provided, calculated, etc.) and prohibit the invention of customer quotes, market share, pricing, or legal claims, which prevents the hallucination of sensitive or misleading information.
- [SAFE]: The skill follows best practices for credential management by explicitly instructing the agent to keep API keys out of skills and source control.
Audit Metadata