m365-onedrive

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata from OneDrive file and folder names which could harbor malicious instructions.
  • Ingestion points: File and folder names retrieved via m365 spo file list and m365 spo folder list.
  • Boundary markers: No delimiters or ignore-instructions are used when listing contents.
  • Capability inventory: The agent has shell access via the Bash tool and can move or delete files.
  • Sanitization: No sanitization of file names is implemented before they enter the agent context.
  • [UNVERIFIABLE_DEPENDENCIES]: Recommends the @pnp/cli-microsoft365 package. This is an established community tool for Microsoft 365 management.
  • [DATA_EXPOSURE]: Fetches account information from Microsoft Graph (graph.microsoft.com). These operations use well-known official endpoints and require an existing authenticated CLI session.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 04:24 AM
Security Audit — agent-trust-hub — m365-onedrive