marketing-claims-review
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- Structured Data Processing: The skill is designed to ingest and analyze untrusted marketing copy provided by the user. While processing external input is a standard surface for indirect prompt injection, the skill mitigates this by applying a strict taxonomy and classification workflow (puffery vs. factual claims) rather than executing instructions within the input.
- Local Configuration and Storage: The skill references and writes to specific paths within the
~/.claude/plugins/config/directory. These operations are used to maintain consistent legal standards and matter-specific context across sessions, which is expected behavior for a tool managing persistent professional workspaces. - External Research Integration: The instructions include steps to research advertising standards via configured tools or web searches. The skill emphasizes the importance of citing primary sources and provides clear attribution tiers for AI-generated content, encouraging human-in-the-loop verification for legal approvals.
Audit Metadata