The Agent Skills Directory

[SAFE]: The skill provides legitimate developer tools for Docker optimization and security auditing. Detailed analysis of the included Python scripts (scripts/dockerfile_analyzer.py and scripts/compose_validator.py) confirms they perform standard static analysis using regex patterns to identify common container anti-patterns and security risks (e.g., running as root, missing healthchecks, or hardcoded secrets).\n- [EXTERNAL_DOWNLOADS]: The skill includes installation instructions fetching from a public GitHub repository (github.com/alirezarezvani/claude-skills.git). This is documented as a standard installation procedure for the skill's components and does not involve executing untrusted code.\n- [COMMAND_EXECUTION]: The skill documents the use of internal Python scripts and standard Docker CLI commands. These are consistent with the skill's primary purpose of container development and are intended for manual execution by the user or as guided steps by the agent.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Dockerfiles and Docker Compose files. While these are untrusted data ingestion points, the skill focuses on static analysis and generating configuration improvements. The presence of analysis logic acts as a functional boundary, and no instructions were found that would cause the agent to execute arbitrary content embedded within the Docker files.

docker-development