skills/teamwiseflow/xiaobei/awk-tts/Gen Agent Trust Hub

awk-tts

Pass

Audited by Gen Agent Trust Hub on Sep 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script uses subprocess.run to invoke ffprobe for audio metadata extraction. The call is implemented using a list of arguments without a shell environment, and it operates on audio files that are generated by the skill itself, which significantly limits the risk of command injection.
  • [EXTERNAL_DOWNLOADS]: When utilizing Alibaba Cloud's Bailian service, the skill downloads generated audio files from the official *.maas.aliyuncs.com endpoint. These downloads are part of the intended functionality and originate from a well-known technology provider.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests text content from external files like tts_requirement.md to perform speech synthesis. While this represents a data ingestion surface, the skill only processes the text for synthesis and similarity comparison, without executing instructions or feeding results back into an autonomous control loop.
  • [DATA_EXFILTRATION]: The skill requires API keys via environment variables to interact with TTS and ASR services. The analysis confirms these credentials are only transmitted to the official API endpoints of the service providers (Volcengine and Alibaba Cloud) for authentication purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 22, 2026, 01:35 PM
Security Audit — agent-trust-hub — awk-tts