The Agent Skills Directory

[SAFE]: The skill provides a legitimate utility for improving Korean text quality by detecting 'translationese' and 'GPT-style' patterns. It does not contain any malicious code or instructions.
[REMOTE_CODE_EXECUTION]: Both scripts (lint.py and export.py) are self-contained and explicitly use only the Python standard library. They do not download or execute remote code.
[DATA_EXFILTRATION]: No network operations (curl, wget, requests) or sensitive file access patterns were found in the scripts or instructions. The scripts process local files provided by the user and output results to the console.
[OBFUSCATION]: Analysis of the regex patterns in patterns.json and the prose in all markdown files revealed no hidden payloads, encoded strings, or obfuscated URLs.
[COMMAND_EXECUTION]: The provided scripts perform safe operations (regex matching and JSON parsing) on input data. They do not use eval(), exec(), or subprocess to run arbitrary commands.
[INDIRECT_PROMPT_INJECTION]: While the lint.py script processes user-supplied text files, it does not execute the content or allow the input to influence the agent's behavior beyond the intended linting functionality. No high-risk capabilities are exposed to the input data.

ko-natural