fabric-ai

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted data from YouTube transcripts (-y ) and web content (-u ), creating an attack surface for indirect prompt injection. * Ingestion points: URL flags in SKILL.md. * Boundary markers: None defined in the prompt instructions. * Capability inventory: Potential for command execution via fabric patterns. * Sanitization: None provided for external inputs.
  • [COMMAND_EXECUTION]: The workflow in SKILL.md includes an 'Execute' step, and several patterns (e.g., 'create_command', 'analyze_threat_report_cmds') are designed to generate CLI commands for security tools or other utilities, which may lead to unauthorized command execution if the agent has shell access.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 12:04 PM