ditto
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process third-party web content, which presents an attack surface for indirect prompt injection.\n
- Ingestion points: External content is introduced through URL fetching and HTML/CSS snippet processing in
SKILL.mdandreferences/sourcing-rules.md.\n - Boundary markers: The instructions establish clear boundaries by stating that fetched content must be treated as data to observe and never as instructions to obey.\n
- Capability inventory: The skill has the ability to implement code via disk writes and perform network operations to fetch source content.\n
- Sanitization: The skill explicitly prohibits executing fetched scripts or following instructions found in comments or metadata.
Audit Metadata