writing-taste

Removes AI writing tells, clichés, and hedging from drafts to restore human editorial voice.

2|Updated Jul 15, 2026
One-click install
npx skills add https://github.com/poorvith-mp/skills-writing --skill writing-taste-poorvith-mp
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: writing-taste
Source: https://github.com/poorvith-mp/skills-writing/tree/main/skills/writing-taste
Command: npx skills add https://github.com/poorvith-mp/skills-writing --skill writing-taste-poorvith-mp

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-generated drafts read as synthetic corporate sludge: uniform rhythm, filler transitions, hedged claims, and exhausted clichés. This Skill audits and rewrites drafts to eliminate those tells and restore concrete, opinionated human prose. ## Core Features & Use Cases - Banned Cliché Registry: Detects and replaces AI-favorite verbs, nouns, adjectives, and structural phrases like "delve", "tapestry", and "in today's fast-paced landscape". - Rhythm & Hedge Repair: Breaks symmetrical triplets, varies sentence length, and strips hedging qualifiers in favor of direct, falsifiable assertions. - Concrete Grounding: Replaces abstract marketing adjectives with specific numbers, benchmarks, and physical nouns. - Use Case: Paste an AI-generated product announcement and receive a DE-SLOP-AUDIT.md report with before-and-after rewrites, slop metrics, and stylistic guidelines. ## Quick Start Ask the assistant to audit this draft for AI writing tells and rewrite it with concrete specifics and varied sentence rhythm.

Frequently Asked Questions about writing-taste

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I remove AI writing tells from a draft?▼

Run the draft through the four-step sanitization protocol: purge banned clichés, break symmetrical triplets, strip hedging qualifiers, and replace abstract adjectives with concrete numbers. The output is a de-slop audit with before-and-after rewrites.

What words and phrases signal AI-generated text?▼

Common tells include verbs like delve, unlock, and harness; nouns like tapestry, beacon, and journey; adjectives like crucial and robust; and openers like "In today's fast-paced landscape." The Skill maintains a banned registry of these terms.

How is this different from line editing or proofreading?▼

Line editing tightens wordy sentences and cuts superfluous words, while this Skill targets generative patterns: clichés, false balance, hedging, and monotonous rhythm. For sentence-level tightening, use a line-editing skill instead.

Can it rewrite corporate marketing copy to sound human?▼

Yes. It replaces hype phrases like "thrilled to announce" and "game-changing" with direct statements grounded in verifiable specifics, such as shipping details, benchmarks, or concrete user impact.

When should I not use this writing style audit?▼

Avoid it for visual design audits, which belong to design-taste, and for extracting a personal voice profile, which belongs to voice-profile. It also does not perform grammar-only proofreading.