azure-storage-blob-ts
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- Trusted Dependencies: The skill utilizes official packages (
@azure/storage-bloband@azure/identity) from a recognized and trusted organization, ensuring the integrity of the underlying code. - Secure Credential Management: The instructions guide users to utilize environment variables for sensitive keys rather than hardcoding them, which is a standard security practice to prevent accidental credential exposure.
- File System Interaction: The skill includes methods for uploading from and downloading to the local file system (e.g.,
uploadFile,downloadToFile). While these are standard SDK capabilities, they should be monitored to ensure the agent only accesses intended paths. - Data Ingestion Surface: The skill enables the agent to retrieve content from external blobs. This creates a potential surface for indirect prompt injection where data stored in the cloud could contain instructions intended to influence the agent's behavior.
- Ingestion points:
blobClient.download(),blobClient.downloadToFile(), andblobClient.downloadToBuffer()inSKILL.mdandreferences/streaming.md. - Boundary markers: None explicitly defined in the provided code snippets.
- Capability inventory: File system writes (
downloadToFile), network operations (SDK API calls), and blob management (CRUD). - Sanitization: The skill focuses on data transport; sanitization of the content itself would occur in the downstream processing logic.
- Least Privilege Access: The skill provides extensive documentation and patterns for generating Shared Access Signatures (SAS) with specific permissions and expirations, supporting the principle of least privilege.
Audit Metadata