modal-container-images

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). Documentation contains explicit examples and APIs that enable credential access/exfiltration and arbitrary code execution during image builds (e.g., copying ~/.aws, run_function/run_commands that fetch and execute remote code, and use of secrets for registry/ECR pulls), which are high-risk primitives that can be abused for data exfiltration and remote code execution.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill contains build-time git clones that fetch and then execute remote code (e.g., "git clone https://github.com/org/repo" followed by "cd /opt/repo && make install"), so these runtime fetches of external repositories can execute uncontrolled remote code — specifically the URLs https://github.com/org/repo and https://github.com/example/repo.git are used.