ado-create-feature
Pass
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill implements a defensive instruction set that explicitly directs the agent to ignore any arguments provided at the command line level. This ensures that only data collected through the controlled interactive flow is used, mitigating potential argument-based injection attacks.
- [SAFE]: The skill's operational flow is restricted to reading local configuration and calling a specific MCP tool.
- Ingestion points: Reads project configuration from the
CLAUDE.mdfile and accepts interactive user input for Feature titles and descriptions. - Boundary markers: Enforces a mandatory interactive requirement and ignores initial command arguments; requires user confirmation for AI-generated content.
- Capability inventory: Uses the
wit_create_work_itemMCP tool for Azure DevOps integration. - Sanitization: Applies HTML formatting to the description field to ensure readability in the Azure DevOps UI; the skill relies on the target platform for backend content sanitization.
Audit Metadata