audit
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it is designed to ingest and analyze external frontend code, pages, and components.\n
- Ingestion points: Processes frontend design context, components, and target pages as specified by user input or workspace files.\n
- Boundary markers: Missing; no explicit boundary delimiters or strict rules are provided to separate the ingested code text from agent instructions.\n
- Capability inventory: Confined to generating structured technical reports and recommending standard project commands; no arbitrary command or shell execution capabilities are included.\n
- Sanitization: Absent; the instructions do not define methods to filter or escape malicious text patterns within the analyzed source code.
Audit Metadata