The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes shell commands such as grep, basename, and wc within its validation section. These are used locally to verify the consistency of the SKILL.md file format, check metadata presence, and ensure no unresolved placeholders remain in the instructions.- [DATA_EXFILTRATION]: There are no network requests, external URL references, or instructions to access sensitive credential files (e.g., .env, .ssh, or AWS configs).- [PROMPT_INJECTION]: The skill contains no instructions designed to override the agent's safety filters or system prompt. It focuses entirely on defining a reusable workflow for skill development.

skill-builder