Proposal: Further hardening container images for MCP

[Mossaka]

Since Agentic Workflows support running custom MCP servers in Docker Containers, we should take advantage of the robust set of security features that container or Open Container Initiative (OCI) provides that are designed to further restrict and secure containerized applications. I'd like to write down a list of features that are in addition to #78 and github/gh-aw-mcpg#275

• Image Signing: verify that the image has not been tampered with since it was signed
• Seccomp: Filter and restrict the system calls a container can make to the host kernel
• AppArmor & SELinux: Use Mandatory Access Control (MAC) profiles to confine container actions, such as file access and network operations.
• Run as Non-Root: The best practice is to run container processes as a non-privileged user
• Linux Capabilities: We can remove all default Linux capabilities and add back only those that are strictly necessary (--cap-drop=ALL)
• Resource Limits (cgroups): Constrain CPU and memory usage to prevent Denial of Service (DoS) attacks

[github-actions]

👋 Auto-assigning this security enhancement issue based on recent contribution activity. This covers important container hardening features like Seccomp, AppArmor, and SELinux for MCP servers running in Docker.

🏴‍☠️ Ahoy! This treasure was crafted by Auto-Assign Issue

[github-actions]

Assigned to @Mossaka since you originally opened this issue and have been actively contributing to container security features in the repository.

Ahoy! This treasure was crafted by 🏴‍☠️ Auto-Assign Issue