skills/microsoft/work-iq/workiq/Gen Agent Trust Hub

workiq

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • Instructional Steering for Tool Usage: The skill includes emphatic instructions such as 'CRITICAL' and 'ALWAYS use WorkIQ' to guide the agent toward using this specific tool for workplace-related queries. While this strongly directs agent behavior, it is intended to ensure the agent leverages the most relevant data source for organizational questions.
  • Sensitive Data Access: The skill is designed to retrieve and summarize information from emails, Teams messages, and internal documents via Microsoft 365 Copilot. This provides the agent with access to high-privilege organizational data, which is necessary for the skill's primary purpose of providing workplace context.
  • Indirect Prompt Injection Surface: Because the skill processes data from external communications (like emails and chat messages), it creates a surface where instructions embedded in those messages could potentially influence the agent's subsequent actions.
  • Ingestion points: Microsoft 365 Copilot data (emails, messages, documents) referenced in SKILL.md.
  • Boundary markers: There are no explicit delimiters or instructions defined in the skill to treat retrieved workplace content as untrusted.
  • Capability inventory: The skill utilizes the ask_work_iq tool to perform queries across the Microsoft 365 environment.
  • Sanitization: No specific validation or sanitization of the retrieved workplace data is mentioned in the prompt configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 10:07 AM