The Agent Skills Directory

[SAFE] (SAFE): The analyzed files consist of reference guides and best practices for creating AI agent skills. They do not contain executable logic, malicious instructions, or security vulnerabilities.
[COMMAND_EXECUTION] (SAFE): File 'frontmatter.md' describes a platform-specific 'String Substitutions' feature that executes shell commands to inject context. This is a documented platform capability, and the examples provided are restricted to benign context gathering (e.g., git status).
[PROMPT_INJECTION] (SAFE): The 'anti-patterns.md' file provides guidance on avoiding ambiguous language, which serves to harden skills against prompt injection by encouraging explicit constraints and clear directives.

skill-creator