x-product-deep-research
Warn
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- COMMAND_EXECUTION (MEDIUM): The skill relies on 'poetry run python' to execute local scripts (deep_research.py and youtube_analyzer.py). Because the source code for these scripts is not included, this constitutes execution of unverified logic within the user environment. \n- DATA_EXPOSURE (LOW): The skill hardcodes and accesses specific local file paths (~/brain/obsidian/Timatron/...) for storing research results and transcripts. Accessing a user's personal Obsidian vault is a sensitive operation. \n- EXTERNAL_DOWNLOADS (LOW): The process involves downloading video content from YouTube and fetching research data from web sources. \n- INDIRECT_PROMPT_INJECTION (LOW): The skill processes research reports and YouTube content which are external and untrusted. \n
- Ingestion points: Research reports and YouTube transcripts processed by scripts. \n
- Boundary markers: None present in the instructions to delimit untrusted content. \n
- Capability inventory: Shell command execution and file system write access. \n
- Sanitization: No sanitization or validation of external data is mentioned before processing.
Audit Metadata