humanizer

Apply statistical humanization levers: vary sentence length between 3 and 35 words, purge overused AI vocabulary like "delve" and "leverage", eliminate em dashes, and replace abstract claims with concrete specifics. A second audit pass catches remaining tells.

Common AI vocabulary markers include "delve", "leverage", "utilize", "robust", "comprehensive", "pivotal", "nuanced", "tapestry", "testament", and "underscores". Structural tells include uniform sentence pacing, em dashes, and forced bold bullet lists.

Yes, the Skill targets documentation, READMEs, tutorials, release notes, PR descriptions, and emails. It preserves technical precision, code snippets, and exact parameters while removing empty buzzwords and filler connectors.

The model that wrote the original draft often misses its own AI tells during a single edit. The Skill requires an explicit second pass asking "What makes this still sound like an LLM?" to catch lingering artifacts like em dashes that sneak back in.

No, the Skill explicitly preserves required technical terms, code snippets, and precise parameters. It only removes empty buzzwords and vague generalizations, replacing them with concrete details like specific latency numbers.