sound-effects-generator

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE] (SAFE): The skill utilizes legitimate Python packages (numpy, scipy, soundfile) for signal processing and audio export. No malicious obfuscation, network exfiltration, or unauthorized credential access patterns were detected.\n- [Indirect Prompt Injection] (LOW): The skill processes external parameters to generate files, creating a potential surface for indirect prompt injection or path traversal.\n
  • Ingestion points: Filenames and paths passed to the .save() method or the --output CLI argument, as well as audio parameters like frequency and duration.\n
  • Boundary markers: None identified. There are no delimiters or instructions provided to the agent to treat input filenames as untrusted data.\n
  • Capability inventory: The skill has the capability to write audio files (WAV, MP3) to the local file system using the soundfile and pydub libraries.\n
  • Sanitization: No evidence of path validation or sanitization was found in the documentation. An agent could be misled into overwriting critical files if it passes an unvalidated user-provided path to the saving functions.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:37 PM