work-iq
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches and executes the
@microsoft/workiqpackage from the official Microsoft scope on the NPM registry vianpxcommands. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from Microsoft 365 sources (emails, messages, documents) which creates an attack surface for embedded instructions.
- Ingestion points: Reads email content, Teams messages, and SharePoint/OneDrive document metadata and content as described in
SKILL.mdandexamples.md. - Boundary markers: The provided instructions do not specify explicit boundary markers or delimiters for separating retrieved external content from the agent's internal logic.
- Capability inventory: The skill is authorized to use
npxandnpmshell commands and invokes the tools provided by theworkiqMCP server. - Sanitization: There are no documented sanitization, filtering, or escaping steps for the content retrieved from the Microsoft 365 API before it is presented to the agent.
Audit Metadata