python
Pass
Audited by Gen Agent Trust Hub on Feb 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The scripts/setup_venv.py and scripts/setup.py files utilize the subprocess module to create Python virtual environments and execute pip install commands to manage dependencies.
- [EXTERNAL_DOWNLOADS]: The skill fetches several Python packages from official registries and uses the pycloudflared library to establish a public tunnel to the local Flask server for receiving VideoDB webhooks.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes unverified video data that may contain malicious instructions.
- Ingestion points: Video and audio content uploaded via URL or file in scripts/batch_upload.py, and transcript data retrieved in scripts/search_and_compile.py.
- Boundary markers: Absent; external data such as transcripts are not wrapped in delimiters or accompanied by instructions to the agent to ignore embedded commands.
- Capability inventory: The skill can execute Python code through the Bash(python:*) tool and perform network operations via the videodb library.
- Sanitization: No sanitization, filtering, or validation is performed on the ingested media content or transcripts before they are presented to the agent context.
Audit Metadata