The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/audit_deployment.py script executes the docker compose config command to analyze the rendered deployment model. This execution is performed using subprocess.run with a list of arguments and without a shell, which is a secure practice that prevents command injection. The tool is used to parse the configuration in a way that respects the deployment's logic.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Dockerfiles and Compose files. While these files are external data that could theoretically contain malicious instructions (indirect injection), the tool only performs static analysis and uses the structured output of the Docker CLI. The agent's instructions focus on interpreting the security findings rather than executing instructions found within the audited files.

harden-deployment