video-publisher
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill orchestrates a pipeline that ingests external data (research results, titles, descriptions, and tags) and passes it to tools with upload and channel management capabilities.
- Ingestion points: Metadata generation fields (Title, Description, Tags) and planning inputs (Topic/niche) defined in the workflow of
SKILL.md. - Boundary markers: The instructions do not define delimiters or specific system instructions for the agent to treat these external data inputs as untrusted content.
- Capability inventory: The workflow delegates tasks to tools such as
youtube-uploaderandyoutube-studiowhich have the capability to modify account state and publish content. - Sanitization: There are no instructions for sanitizing, validating, or escaping user-provided or research-derived metadata before it is interpolated into the publishing process.
- [NO_CODE]: The skill file consists entirely of documentation and workflow orchestration instructions in markdown format, with no scripts, binary code, or executable components embedded.
Audit Metadata