microsoft-onedrive

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill communicates with official Microsoft Graph API endpoints (graph.microsoft.com) for all operations. As a well-known service, these network references are consistent with the skill's stated purpose.
  • [COMMAND_EXECUTION]: Standard utilities such as curl and jq are employed to interact with the API. The skill provides structured recipes for common tasks like listing, searching, and downloading files, utilizing proper quoting for variables to prevent shell injection.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external data from OneDrive and SharePoint, which introduces a surface for indirect prompt injection.
  • Ingestion points: Downloads of file content and item metadata from graph.microsoft.com (SKILL.md).
  • Boundary markers: Boundary markers for external file content are not defined within the interaction recipes, which is expected for raw file handling.
  • Capability inventory: Includes shell execution (Bash), file write access to the skill's temporary directory, and network connectivity (SKILL.md).
  • Sanitization: No specific sanitization methods are described for the data retrieved from external storage, as the agent is intended to process the files as-is.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 05:12 PM