landing-page-design

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate landing-page design guide, but its top-line install directive points to a belt-sh/cli package and a raw GitHub install script, while the tool it describes is inference.sh's, so the source is unverified and risky toward the user without clear malicious intent.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). The required workflow runs tavily/search-assistant (lines 28-30) and exa/answer (lines 171-173), which the agent must actively invoke to retrieve arbitrary web content, so outsider-authored pages are reached only through an agent-initiated search (low). The image-generation and html-to-image calls send only the agent's own prompts and do not ingest outsider text.