How I Turned AI Writing Tells into a Pre-Read Quality Gate

Description and use

Implementing the humanizer gate lets teams score AI drafts before editing, saving editor time and improving content quality.

Problem

Editors waste time polishing AI-generated drafts that contain recognizable machine‑writing patterns.

Outcome

Implementing the humanizer gate reduced editor time spent on mechanical edits by allowing early identification of low‑quality AI drafts.

Workflow

• 1. Load the humanizer.md skill into the AI agent before writing.
• 2. The agent starts a draft score at 100 and subtracts points for each detected AI‑writing pattern.
• 3. Repeated patterns incur up to twice the base penalty.
• 4. The agent reports the final score to the team before the draft is reviewed.
• 5. If score <50, rewrite the draft; if 50–70, target high‑impact patterns; if >70, proceed to full editorial review.

Evidence

The skill lives in a single Markdown file, humanizer.md, inside a content ops folder in my open-source ai‑marketing‑skills repo.

Source

At a high level, the process is simple: Start every draft at 100. Subtract points for each AI‑writing pattern. Stack repeated uses of the same pattern, up to twice the base penalty. Read the score before anyone reads the draft.

Source

The rubric sorts results into four bands, and one of those bands is blunt: anything under 50 “Reads like ChatGPT output.”

Source

The file is written for Claude Code, Anthropic’s coding agent, but the setup is portable as long as the agent can read a file.

Source

The rubric is based on Wikipedia’s public page on Signs of AI writing, which lists 24 scored patterns.

Source