The Agent Skills Directory

[PROMPT_INJECTION]: No attempts to override safety guidelines or extract system prompts were found. The instructions focus entirely on the stated purpose of text editing and humanizing content.
[DATA_EXFILTRATION]: No network operations, sensitive file access patterns, or credential harvesting techniques were detected.
[REMOTE_CODE_EXECUTION]: No external dependencies, package installations (pip/npm), or remote script execution patterns are present in the skill files.
[COMMAND_EXECUTION]: The skill requests standard file manipulation tools (Read, Write, Edit, Grep, Glob) which are appropriate for its purpose. No suspicious shell commands or privilege escalation attempts were found.
[OBFUSCATION]: No hidden or encoded content such as Base64, hex, or zero-width characters was detected in the provided files.
[DATA_INJECTION_SURFACE]: While the skill processes user-provided text, which is an inherent surface for indirect prompt injection, it focuses on text editing/rewriting and does not incorporate capabilities that would allow such input to trigger dangerous actions.

humanizer