marketing-page-layout
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's
AUDITfunctionality allows it to fetch and parse content from arbitrary external URLs. This creates a surface for indirect prompt injection, where an attacker could place hidden instructions on a website to manipulate the agent's evaluation or subsequent output.\n- Ingestion points:SKILL.md(AUDIT mode) uses tools like browser MCP, Playwright, or Puppeteer to ingest DOM content from external sites.\n- Boundary markers: The skill lacks explicit instructions to isolate or treat external content as potentially untrusted data.\n- Capability inventory: The skill can generate technical specifications and static code (BUILD mode), making it a potential target for manipulation.\n- Sanitization: There is no evidence of content sanitization or instruction to ignore embedded commands within the audited data.
Audit Metadata