channel-teardown
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from YouTube video titles and transcripts, which could potentially contain malicious instructions intended to influence agent behavior.\n
- Ingestion points: Data retrieved from
get_channel(titles) andget_video_transcript(transcripts).\n - Boundary markers: The skill uses a structured output format for its diagnosis essay but does not explicitly define delimiters for isolated processing of external content.\n
- Capability inventory: Execution is restricted to TubeLab MCP tool calls; the skill lacks access to shell commands, the local file system, or arbitrary network endpoints.\n
- Sanitization: There are no instructions for sanitizing or escaping the content retrieved from YouTube before it is incorporated into the analysis report.
Audit Metadata