piper-tts-training
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill documentation and accompanying scripts are focused on the task of training text-to-speech models and do not contain any malicious patterns.
- [INDIRECT_PROMPT_INJECTION]: The workflow involves processing external text corpora for voice training.
- Ingestion points: The metadata.csv file and raw text corpus files are read by the convert_spelling.py script and the piper_train.preprocess module.
- Boundary markers: The skill does not explicitly define delimiters for untrusted corpus text.
- Capability inventory: The skill utilizes Read, Write, and Bash tools to manipulate files and execute training processes.
- Sanitization: No explicit sanitization or filtering of the corpus text is performed before it is passed to the phonemization and training modules.
Audit Metadata