yt-seo
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructions involve reading from and writing to a local file path at
~/.claude/youtube/voice.md. While this is used for personalizing the agent's output, it demonstrates access to user-specific configuration data. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data when it asks the user to provide videos for voice inference.
- Ingestion points: Transcripts or content from three user-provided videos (SKILL.md).
- Boundary markers: None. There are no delimiters or instructions to ignore embedded commands within the video content.
- Capability inventory: The skill has file-writing capabilities to the local filesystem (
~/.claude/youtube/voice.md). - Sanitization: There is no evidence of sanitization or filtering of the video content before it is processed by the agent.
Audit Metadata