fabric

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests and processes untrusted public content (e.g., "curl -s https://example.com/article | fabric -p summarize" in Step 2 and "fabric -y 'https://youtube.com/watch?v=VIDEO_ID'" in Step 4), meaning the agent will read and act on arbitrary web/YouTube content which could contain instructions that influence subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches and uses external pattern files from the Fabric GitHub repo at runtime (e.g., fabric -u / Pattern Library: https://github.com/danielmiessler/Fabric/tree/main/patterns), which directly control agent prompts, and the installer command also downloads-and-executes remote code via https://raw.githubusercontent.com/danielmiessler/fabric/main/scripts/installer/install.sh | bash — both are runtime/setup fetches that can change prompts or execute code.