The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file paths, or unauthorized network operations were detected. The templates include guidance for .dockerignore files to prevent the accidental inclusion of sensitive local environment data or build artifacts.
[Unverifiable Dependencies] (SAFE): All container images referenced (mcr.microsoft.com) are from Microsoft's official, trusted registry. No untrusted third-party scripts or packages are used.
[Privilege Escalation] (SAFE): The skill promotes security hardening by recommending the USER app directive and 'chiseled' images which lack a shell and run as non-root by default, preventing common container escape and post-exploitation techniques.
[Indirect Prompt Injection] (SAFE): The skill ingests project metadata (e.g., from .csproj files) to configure the build process.
Ingestion points: Project files (.csproj) and user version input.
Boundary markers: None (standard template generation).
Capability inventory: No script execution or network operations in the skill itself; output is static text templates.
Sanitization: Not applicable as the output is a text file for developer review.

dotnet-dockerfile