b2-backup-restore
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill adheres to the principle of least privilege and data isolation by explicitly prohibiting the transfer of object data or presigned URLs through the AI model's context or chat interface.
- [INDIRECT_PROMPT_INJECTION]: The skill interacts with external data sources such as B2 bucket listings and object metadata (ingestion points in
SKILL.md). It identifies and mitigates the risk of indirect prompt injection by requiring data to be handled out-of-band (structural boundary), utilizing specific MCP tools for orchestration (capability inventory inSKILL.md), and strictly prohibiting the display of object contents in chat (sanitization). - [DATA_EXPOSURE_EXFILTRATION]: Instructions are provided to handle sensitive presigned URLs as bearer material, ensuring they are only passed to out-of-band clients and never exposed within the chat session or routed through the server.
- [COMMAND_EXECUTION]: The skill uses a defined set of B2 and S3 API tools for storage orchestration and does not include patterns for arbitrary command execution or unauthorized system access.
Audit Metadata