paw-pa-intake
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill interacts with the AssemblyAI API (api.assemblyai.com) to upload local media files and retrieve transcripts. It also allows the agent to process content from user-provided URLs for transcription purposes.
- [COMMAND_EXECUTION]: The workflow involves running local Python scripts (scripts/transcribe.py and scripts/write_brief.py) to automate the transcription process and the generation of structured brief files. The test scripts also utilize dynamic importing to load these local modules for unit testing, which is standard practice.
- [PROMPT_INJECTION]: As an intake skill processing untrusted multimodal data (audio, video, and text), there is an inherent risk of indirect prompt injection where adversarial instructions could be embedded in the source material. The skill mitigates this by emphasizing structured extraction, validation, and synthesis rather than verbatim output of the processed transcript.
Audit Metadata