customer-service-pro
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the design of systems intended to handle untrusted customer input across various social messaging platforms. While the skill generates static documentation, the resulting customer service systems have an inherent exposure to indirect prompt injection vectors.
- Ingestion points: Customer messages received via Line OA, Facebook Inbox, IG DM, and TikTok Shop, as outlined in SKILL.md and templates/output-template.md.
- Boundary markers: The provided output templates lack explicit instructions or delimiters designed to isolate user-provided data from agent instructions in the final implementation.
- Capability inventory: The skill is restricted to generating Markdown content and does not involve subprocess execution, network requests, or direct file system modifications.
- Sanitization: The skill provides high-level guidance on privacy (PDPA) and professional response standards but does not include technical implementation details for input validation or sanitization in the generated chatbot logic.
Audit Metadata