landing-page-design

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). Piping a remote shell script from cli.inference.sh (and related inference.sh) into sh is a high-risk distribution pattern — it's an unverified .sh installer from a non-obvious vendor and could easily be used to distribute malware unless you can fully verify the publisher and inspect the script first.

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md Quick Start and Research sections explicitly run external search/QA apps (e.g., infsh app run tavily/search-assistant and infsh app run exa/answer) that fetch and ingest public web pages/search results for competitor research and social-proof stats, so the agent would read untrusted third-party content that could influence its decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). The Quick Start includes a runtime command that downloads and pipes remote code to a shell (curl -fsSL https://cli.inference.sh | sh), which executes remote code and is required for the skill’s infsh commands that run models, so this external URL directly controls execution.