The Agent Skills Directory

[SAFE]: The skill is composed of markdown and JSON files containing only text-based instructions and test scenarios. No malicious behavior or security vulnerabilities were identified.\n- [NO_CODE]: No source code, scripts, or binary executables are included in this skill. The package consists entirely of documentation and test cases.\n- [PROMPT_INJECTION]: The skill content does not contain any instructions intended to bypass safety filters, override core system prompts, or extract sensitive information from the model.\n- [DATA_EXFILTRATION]: There are no network requests, file access patterns, or credential harvesting techniques present in the skill files.

liking