ms-onedrive-personal-graph

Pass

Audited by Gen Agent Trust Hub on Oct 25, 2026

Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill stores sensitive OAuth tokens in the hidden local file path ~/.onedrive-mcp/credentials.json. Access to this path by other scripts or processes could lead to credential exposure.
  • [COMMAND_EXECUTION]: The skill relies on shell scripts (onedrive-setup.sh, onedrive-cli.sh) to perform operations, utilizing system tools such as bash, curl, and jq for network communication and data parsing.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data through OneDrive file paths and folder names as arguments to onedrive-cli.sh. Ingestion points: Path and filename arguments in SKILL.md. Boundary markers: Absent. Capability inventory: bash, curl, jq, and python3. Sanitization: Not specified in the provided instructions, creating a risk if input is not escaped before shell execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 25, 2026, 07:02 AM