The Agent Skills Directory

[SAFE]: The skill's primary function is to serve as a document builder, using a variety of predefined Markdown templates in the references/ directory. It instructs the agent to transform unstructured user input into professional artifacts without performing dangerous operations.
[NO_CODE]: There are no scripts (Python, Node.js, or Bash) included in the skill. All logic is contained within declarative instructions and static templates, eliminating the risk of runtime code execution or library-based attacks.
[COMMAND_EXECUTION]: The reference file references/general/agent-skill-architecture.md mentions and demonstrates shell command injection syntax (e.g., !pwd). This is included strictly as documentation and instructional content for users interested in building other skills, and it does not represent an active or exploitable command execution path in this skill.
[PROMPT_INJECTION]: The skill includes reference materials that utilize persona-based role-play (e.g., 'You are an expert AI Engineer'). These are intended as templates for the agent to follow when performing specific architectural tasks and do not attempt to override the platform's safety guidelines or bypass instructions.
[SAFE]: The assets/general/excalidraw-template.md file contains LZ-string compressed data used for Obsidian Excalidraw diagrams. This is a standard configuration format for the plugin and does not contain hidden commands or malicious payloads.

kernel-write-doc