Verified project record

conorbronsdon/avoid-ai-writing

A structured skill and deterministic detection engine that audits text for identifiable AI writing patterns and rewrites content to remove them. It provides transparent reporting on detected patterns across severity levels and includes tiered vocabulary replacements.

Project overview

Provides a zero-dependency JavaScript detection engine and a structured audit workflow with 53 pattern categories, distinguishing it from a single-prompt approach.

Project type: AI Agent · Data Processing · Prompt Engineering
Use cases: Documents & Office
Deployment: Refer to project documentation
License: MIT

Best for

Creators and general users seeking a structured, deterministic tool to audit and rewrite text for AI writing patterns.
Developers needing a zero-dependency detection engine that can be installed via source or required directly in Node.js without a build step.

Key capabilities

Flags AI writing patterns and rewrites the text to fix them, including a second pass to catch surviving patterns.
Flags AI writing patterns without rewriting, showing which flags are real problems versus judgment calls.
Edits a file in place with minimal, targeted changes, preserving passages that are already human, and returns an edits-made report.
Optional voice profiles (casual, professional, technical, warm, blunt) that set how the resulting prose should sound.
An 111-entry word replacement table across 3 tiers plus 10 Tier 3 phrases for identifying and replacing AI vocabulary.

Limitations and risks

Conversational-register patterns are judgment calls rather than regex-detectable, meaning an ordinary human paragraph and an AI-generated one can look structurally identical.
Telemetry tracking, end-to-end data boundaries, cost dependencies, and interaction modes are not documented. Coding requirements are optional.

Getting started

Installation is rated as easy due to a zero-dependency engine with no build step. Users can install via git clone or file copy for agent skills, or require the module directly in Node.js.

Alternatives and comparisons

A zero-shot Text-to-SQL approach using Monte Carlo Tree Search to iteratively infer SQL construction actions and generate accurate queries.
A collection of investment research skills applying four value investing masters' methodologies via multi-agent adversarial analysis to generate actionable investment reports.
A curated repository of AI-powered trading strategies, quantitative dashboards, multi-agent backtesting tools, and investment columns offering verifiable reasoning trails.

Project comparisons

Evidence and sources

README: Audit & rewrite content to remove AI writing patterns. A practical skill for any AI agent. Supports detect-only and edit-in-place modes, plus voice profiles.
GitHub project description: Skill that audits and rewrites content to remove AI writing patterns. Use it with your favorite agents including Claude Code, OpenClaw, and Hermes.
README: Audits and rewrites content to remove AI writing patterns ("AI-isms").
README: A one-shot "make this sound human" prompt catches the obvious stuff. This skill is different: - Structured audit — returns identified issues with quoted text, the rewrite, a chang…
README: **Works across platforms** — one `SKILL.md` runs in Claude Code, Cowork (as a plugin), OpenClaw, and Cursor (as a ported rule). See the install paths below.

Reviewed at: 2026-07-20T13:28:01Z

Open original GitHub project