The Agent Skills Directory

Verified Command Execution: The skill performs informational shell commands such as az account show and git rev-parse to retrieve environment context and repository metadata. These actions are standard for deployment-related tasks and are restricted to read-only or session-specific operations.
Trusted Tooling Maintenance: It utilizes specialized MCP tools to ensure required CLI extensions are installed or updated. This automated setup facilitates the necessary environment for Azure interactions through official, vendor-supported channels.
Static Analysis Protection: The skill prioritizes static analysis of repository files over code execution. It explicitly prohibits running package manager commands like npm install or pip install on user repositories unless high-level remediation or scaffolding is performed with explicit, per-command user consent.
Vulnerability Screening: A dedicated logic layer scans for 'Intentionally Vulnerable Applications' (e.g., security labs) by searching for specific code structures and metadata flags. If such patterns are detected, the skill halts the pipeline to prevent the accidental deployment of insecure software.
Secure Session Management: Application state is managed through locally scoped session directories using unique UUIDs. The skill enforces path-scoping rules to ensure that session artifacts are written only to designated subdirectories, mitigating risks associated with unauthorized file access or directory traversal.

azure-app-onboard-prereq