ai-content-review

To detect AI-generated tells without changing your writing style, the skill surfaces exact text offsets and category flags for AI-like patterns while preserving your historical voice. It provides a read-only, non-destructive diff rather than automatically rewriting your content.

The best way to flag AI writing patterns in emails and code comments is to use context-aware detection that adjusts sensitivity by content type and register. This ensures accurate identification of AI tells across various document formats without altering the original structure.

You can get a structured output report for AI text review to use in automated workflows. The skill produces a machine-friendly report with verdict-ready fields, including exact offsets, categories, and suggested removals for seamless downstream processing.

AI tell detection works on existing documents without modifying the source text. The review process is strictly read-only and non-destructive, returning a diff and flags that you can choose to apply manually if you request a cleaned version.

Stylometry helps preserve voice when removing AI features from articles by analyzing your historical writing style. The skill respects this established voice and adjusts its flagging sensitivity based on the specific content type and register being reviewed.

A limitation when using websearch verification for AI content review is that it focuses strictly on surfacing patterns rather than rewriting text. It relies on a live tells catalog to fetch current patterns, meaning offline environments cannot update the detection categories.