m365-admin

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface detected.
  • Ingestion points: The skill reads untrusted data from Microsoft 365 sources, such as message content and email bodies (e.g., getInboxMessages in scripts/setup_exchange.ts and listMembers in scripts/configure_teams.ts).
  • Capability inventory: It has high-privilege access to M365 identity and resource management and can execute shell commands via the Bash tool.
  • Boundary markers: There are no boundary markers or instructions used to prevent the agent from following instructions embedded in the retrieved data.
  • Sanitization: Data fetched from M365 is not sanitized or validated before being handled by the agent logic.
  • [COMMAND_EXECUTION]: Authorization for shell command execution.
  • The skill is configured to use the Bash tool to run PowerShell and M365 CLI commands. This grants the agent direct interaction with the underlying system shell while holding sensitive cloud credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 08:31 AM
Security Audit — agent-trust-hub — m365-admin