arabic-transcript-cleanup
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a set of instructional guidelines for an AI agent to process Arabic text transcripts. It focuses on correcting dialect-specific errors and marking filler words.
- [EXTERNAL_DOWNLOADS]: The skill mentions standard tools like OpenAI Whisper and provides a command to install
faster-whispervia pip. Both are well-known and legitimate tools for speech-to-text processing. - [COMMAND_EXECUTION]: The provided commands are for local environment setup and usage of standard transcription software. They do not involve privilege escalation or dangerous operations.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes external transcript data, it does not possess capabilities (such as network access or file system modification) that could be exploited via malicious content embedded in the transcripts. The risk is negligible.
Audit Metadata