add-ai-webapi
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEDYNAMIC_CONTEXT_INJECTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- Dynamic Context Injection: The skill uses a shell command placeholder (
!node) to run a local version check script when the instructions are loaded. This is used for environment verification and does not involve network exfiltration or sensitive file access. - Command Execution: The skill performs standard development operations, including executing
npm run buildto verify code compilation and usinggitcommands for version control staging and commits. These are limited to the project directory and specific files modified by the skill. - Indirect Prompt Injection Surface: As a code integration tool, the skill reads existing project files (such as
package.json,src/, and Power Pages configuration YAMLs) to determine where to insert AI features. While this is an attack surface common to static analysis tools, the skill provides specific instructions for human review and includes a validator script to catch potential issues in the generated output. - External Content Access: The skill is designed to fetch documentation and guidelines from Microsoft's official developer resources (
learn.microsoft.com). These references are used to ensure the generated code aligns with the latest service shapes and requirements. - Security Validation Script: The skill includes a Node.js script (
validate-ai-webapi.js) that performs static analysis on the generated code. It enforces security requirements such as mandatory CSRF tokens (__RequestVerificationToken), specific OData headers, and proper Content-Type handling for the target APIs.
Audit Metadata