audio-understanding

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill instructions use npx -y superconductor-gemini-skills to download and execute an external package. This package is not from a trusted organization or repository listed in the security guidelines.
  • REMOTE_CODE_EXECUTION (MEDIUM): The use of npx -y allows for the execution of remote code at runtime. Without specific version pinning or source verification, this presents a supply chain risk.
  • COMMAND_EXECUTION (LOW): The skill executes shell scripts (gemini.sh) and system commands (npx) to facilitate audio processing. While necessary for the skill's function, it increases the attack surface.
  • PROMPT_INJECTION (LOW): The skill is vulnerable to indirect prompt injection via transcribed audio content.
  • Ingestion points: Local audio files and YouTube URLs processed via the --file argument.
  • Boundary markers: Absent; there are no delimiters or instructions to ignore embedded commands in the transcribed text.
  • Capability inventory: The skill executes shell commands and interacts with the Gemini API to process and summarize data.
  • Sanitization: None; the skill transcribes and directly processes external content without validation or escaping.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:44 PM