yt-ideation
Warn
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The sections titled Overview, Prerequisites, Instructions, Output, and Error Handling are entirely unrelated to the skill's primary purpose of YouTube ideation. They reference 'ORM configuration', 'ORM concepts', and 'ORM environment', indicating a potential copy-paste error or misleading documentation that obscures the skill's actual function.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions specify loading external context from 'niche-analysis.json' and 'niche-report.md'. This creates an ingestion surface for untrusted data that could influence the agent's behavior without explicit boundary markers or sanitization steps mentioned in the instructions.
- [CREDENTIALS_UNSAFE]: The error handling table mentions 'Authentication failure' and suggests the user 'Refresh tokens or re-authenticate with ORM'. While this appears to be part of the mismatched ORM metadata, it introduces unnecessary references to credential management and authentication flows in a tool intended for content brainstorming.
Audit Metadata