The Agent Skills Directory

[SAFE]: The skill demonstrates a strong security posture by mandating a local validation step (mvn validate) before processing, ensuring that the agent works within a stable and verified project environment.\n- [SAFE]: It explicitly addresses supply chain security by restricting PlantUML to local includes and prohibiting the retrieval of remote container images, which prevents unauthorized remote code execution or dependency confusion attacks.\n- [COMMAND_EXECUTION]: The skill uses the mvn and plantuml CLI tools to perform its tasks. These commands are standard for Java development and are invoked in a manner consistent with generating documentation from local source code.\n- [PROMPT_INJECTION]: The skill utilizes a highly structured, step-based interactive question flow that limits the scope of the agent's actions and reduces the likelihood of instruction bypass through user input.\n- [DATA_EXFILTRATION]: While the skill analyzes source code and configuration files, it does not contain logic for exfiltrating sensitive data to external servers, focusing instead on producing local diagram files.

033-architecture-diagrams