youtube-channel-search
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from YouTube, which could contain malicious instructions meant to manipulate the agent.
- Ingestion points: The script fetches external content including video titles, descriptions, and transcript text from YouTube using
yt-dlpand a sibling script (scripts/youtube-channel-search.py). - Boundary markers: There are no boundary markers or explicit instructions provided to the agent to treat the fetched content as data rather than instructions.
- Capability inventory: The skill is capable of writing search results to the filesystem and executing system commands.
- Sanitization: The script performs lexical scoring but lacks sanitization or filtering to detect or neutralize prompt injection patterns within the YouTube content.
- [COMMAND_EXECUTION]: The script executes external binaries and sibling scripts to perform cataloging and transcript fetching.
- Evidence: The script
scripts/youtube-channel-search.pyusessubprocess.runto callyt-dlpand a sibling transcript script. - Context: The implementation is secure as it uses argument lists and avoids
shell=True, preventing common command injection vulnerabilities despite handling user-supplied URLs.
Audit Metadata