review-hub-feedback
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes repository comments and feedback, which are external and untrusted sources. \n- Ingestion points: Data is retrieved via
list_repository_commentsandget_repository_comment_threadas specified inSKILL.md. \n- Boundary markers: Instructions explicitly require labeling ingested data as untrusted, using JSON serialization for data structures, and ignoring any embedded commands or tool calls. \n- Capability inventory: The skill writes findings to private session artifacts and can invoke theprepare-feedback-proposalsskill based on user request. \n- Sanitization: Redaction of credentials, secrets, and personal identifying information (PII) is mandatory before evidence is saved or shared between contexts.
Audit Metadata