landing-page-design

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). These URLs point to a non-standard domain that serves an install script (curl | sh) which fetches platform binaries from the site's own dist subdomain and verifies checksums hosted on the same domain — a common but risky distribution pattern because the domain could be compromised and there's no independent signature or well-established reputational signals.

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow includes using the tavily/search-assistant app (via infsh app run tavily/search-assistant --input ...), which fetches public web content and feeds the retrieved page text/snippets into the agent’s LLM context, creating an outsider free-text injection surface.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The Quick Start invokes remote code execution via "curl -fsSL https://cli.inference.sh | sh" (which then downloads binaries from https://dist.inference.sh) to install the infsh CLI, and the skill requires that CLI at runtime, so the URLs directly provide and execute remote code.