The Agent Skills Directory

[SAFE]: Analysis of the skill reveals no signs of malicious intent or hidden patterns. The tool is designed for the specific, benign purpose of text encoding repair.\n- [REMOTE_CODE_EXECUTION]: The skill does not perform any remote code execution or external downloads. All functionality is contained within the provided Python script, which uses only the standard library.\n- [INDIRECT_PROMPT_INJECTION]: The tool operates on external text data provided to the sanitize_outbound function. While this data could theoretically contain malicious instructions, the skill lacks dangerous capabilities such as network access, file system modification, or command execution that could be exploited by such instructions. The ingestion point is limited to string processing within the Python runtime.\n- [EXTERNAL_DOWNLOADS]: The skill's provenance refers to an external repository on GitHub under the author's organization. This reference is for source tracking and does not result in runtime downloads or installation of untrusted content.

encoding-fix