fluent-2

Pass

Audited by Gen Agent Trust Hub on Sep 27, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references several official Microsoft design resources (e.g., fluent2.microsoft.design) and the official Fluent UI GitHub repository (github.com/microsoft/fluentui). These links are used as runtime references to guide the agent on component roles, color systems, typography, and motion principles. These references point to official documentation and are not associated with any malicious content.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process descriptions of application tasks and information architecture to translate them into Fluent 2 design tokens. While it processes external data, the skill lacks capabilities such as shell command execution, file system writes, or network exfiltration, which significantly limits the attack surface for prompt injection.
  • [COMMAND_EXECUTION]: Although the skill body contains a code block mentioning 'Skills CLI discovery / standalone install', it is provided as a text instruction for manual verification by a developer and does not contain any executable scripts or automated command triggers.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 27, 2026, 12:48 PM