avoid-ai-writing
Audit and rewrite prose so it stops reading as machine-generated
by wshobsonFree
Description
avoid-ai-writing is an agent skill from the wshobson/agents collection that audits prose for the patterns that make writing read as machine-generated, and then fixes them. It finds those patterns, commonly called AI-isms, and rewrites them while trying to keep the author's own voice rather than sanding every sentence into the same tone. The skill is careful about what its flags mean. Its guidance says that these patterns also appear in human writing, especially under deadline, in an unfamiliar genre or in a second language, and that detection tools give unreliable results, so every flag is treated as a writing-quality signal rather than proof of who wrote something. It offers three modes. The default rewrite mode flags the patterns, returns a clean version with the editable AI-isms removed, and summarizes what changed. A detect mode only flags patterns and says which are clear problems and which are judgment calls, which suits published or someone else's text. An edit mode changes a prose file in place, and the skill refuses source code or configuration files. It is installed with the skills CLI from the wshobson/agents repository, and its skills.sh page showed passing results from three security audits at the time of the scan.…
Highlights
- Finds and rewrites AI-ism patterns in prose
- Keeps the author's voice
- Rewrite, detect and in-place edit modes
- Separates clear problems from judgment calls
- Refuses code and configuration files in edit mode
- Installs with the skills CLI
Ratings & Reviews
No ratings yet
Be the first to rate avoid-ai-writing.
People also visit
canvas-design
Create beautiful visual art in .png and .pdf documents using design philosophy
avoid-ai-writing FAQ
What does this skill do?
It finds patterns that make text read as machine-generated and rewrites them while keeping the author's voice.
Does it tell me whether text was written by AI?
No. The skill says these flags are writing-quality signals, not proof of authorship, and they should not decide academic, hiring or attribution questions.
What is the difference between the modes?
Rewrite, the default, returns a cleaned version. Detect only flags patterns. Edit changes a prose file in place.
Will it edit my code files?
No. In edit mode the skill refuses source code and configuration files and only works on prose.
When is detect mode better?
When you want to decide the changes yourself, or when the text is published or belongs to someone else.
How do I install it?
Run npx skills add https://github.com/wshobson/agents --skill avoid-ai-writing.
What security audits does it have?
skills.sh lists three audits: Gen Agent Trust Hub Pass, Socket Pass and Snyk Pass.
What are the best avoid-ai-writing alternatives?
Popular alternatives to avoid-ai-writing include implement, brand-guidelines and canvas-design. See all avoid-ai-writing alternatives