The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and humanize user-provided content, which acts as a primary ingestion point for external data. While this theoretically allows for adversarial instructions to be presented to the agent, the skill's instructions strictly guide the agent to perform auditing and rewriting tasks based on specific linguistic checklists, providing a safe operational framework.
[COMMAND_EXECUTION]: The skill invokes a local script, scripts/humanizer_scorer.py, to calculate humanity scores for text. Analysis of the script confirms it is safe, utilizing standard Python libraries (sys, re, json, math) to perform regex-based pattern matching. It does not possess capabilities for network communication, persistent storage, or access to sensitive system directories.

content-humanizer