paper-animation-ad
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches and processes data from external product pages and websites to build product context. This ingestion of untrusted data represents a potential attack surface for indirect prompt injection.
- Ingestion points:
SKILL.mdinstructs the agent to fetch product page URLs and research company websites. - Boundary markers: The referenced
references/agent-security.mdfile contains explicit instructions to treat all fetched content as data, never as instructions, and to report any detected injection attempts. - Capability inventory: The skill is authorized to read/write to the
.agents/directory and generate prompts for external tools. It does not demonstrate high-privilege shell or network execution capabilities. - Sanitization: The instructions require the agent to quote injection attempts as findings rather than executing them, preventing silent obedience to embedded commands.
Audit Metadata