The Agent Skills Directory

[SAFE]: The skill is primarily instructional, providing markdown templates and JSON schemas to guide the creation of standardized skills. It does not contain any executable code or directives that attempt to bypass safety protocols.\n- [COMMAND_EXECUTION]: Reference documents mention that skills may include a scripts/ directory with TypeScript files executed via pnpx tsx. This is presented as a structural guideline for developers and does not involve the execution of malicious or unverified commands by the skill itself.\n- [DATA_EXFILTRATION]: Analysis confirmed the absence of network-related commands (e.g., curl, wget), hardcoded credentials, or access to sensitive local environment paths. The skill's operations are confined to documentation and local file structure management.\n- [PROMPT_INJECTION]: The skill's frontmatter and instructions do not include override commands, safety bypasses, or requests to extract system-level prompts. All instructions are focused on the intended purpose of skill development.

create-skill