resonance-studio

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • Indirect Prompt Injection (HIGH): The skill is highly vulnerable to indirect injection as it processes untrusted user instructions to drive high-privilege tool usage.
  • Ingestion points: User-provided asset 'briefs' defined in the 'Jobs to Be Done' section of SKILL.md.
  • Boundary markers: None; there are no instructions to delimit user input from system instructions or to ignore embedded commands.
  • Capability inventory: read_file, write_file, and generate_image tools are active (SKILL.md).
  • Sanitization: None; user-supplied project names and subjects are directly recommended for use in constructing file paths and names in the 'Asset Generation Pipeline' (references/asset_generation_pipeline.md).
  • Command Execution (MEDIUM): The skill's workflow explicitly utilizes the write_file tool to save generated assets. Without path sanitization or validation, an attacker could attempt to use directory traversal sequences (e.g., '../../') within user-controlled fields like 'Project' to write files to unauthorized directories on the host system.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 09:12 AM