New: Earn 20% as an AgentAlley affiliate. Share your link and earn on every sale you refer. Join now New: Earn 20% as an affiliate. Join now

Writes style - no AI slop

Writing Style - No AI Slop

by swan-gtm · MIT license · ★ 159 on GitHub · updated Sep 23

Writing Style - No AI Slop cover

Use this skill for the writing bar on everything ads-related — operator voice, no AI slop, banned words, and sentence-level rules.

---
name: "ads-writing-style"
title: "Writing style: no AI slop"
description: "Use this skill for the writing bar on everything ads-related — operator voice, no AI slop, banned words, and sentence-level rules."
category: Ads
---

# Writing Style - No AI Slop

This governs everything this agent writes: chat replies, ad copy, audits, plans, reports. The goal is to sound like a sharp operator, not a language model. If a sentence could appear in any generic marketing blog, rewrite it.

## The one test
Read it out loud. If it sounds like a human who knows the work would say it, keep it. If it sounds like a press release or a LinkedIn motivational post, kill it.

## Banned words and phrases
Never use these. They are the fingerprints of AI writing:

- delve, dive in, deep dive, unpack, navigate, landscape, realm, tapestry, journey

Overview

This governs everything this agent writes: chat replies, ad copy, audits, plans, reports. The goal is to sound like a sharp operator, not a language model. If a sentence could appear in any generic marketing blog, rewrite it.

Manual install (npx, ZIP)
Claude Codenpx skills add swan-gtm/gtm-skills --skill ads-writing-style -g -a claude-code
Claude Code (manual)mkdir -p ~/.claude/skills && curl -sL https://codeload.github.com/swan-gtm/gtm-skills/tar.gz/e65d97205fd2 \ | tar -xz -C ~/.claude/skills --strip-components=3 gtm-skills-e65d97205fd2/skills/ivan-falco/ads-writing-style
Windows (PowerShell)New-Item -ItemType Directory -Force "$HOME\.claude\skills" | Out-Null; Invoke-WebRequest "https://m.agentalley.io/zip/1bd88737adb8080ee661/ads-writing-style.zip" -OutFile "$env:TEMP\ads-writing-style.zip"; Expand-Archive "$env:TEMP\ads-writing-style.zip" -DestinationPath "$HOME\.claude\skills" -Force
Claude.ai / desktopDownload the ZIP and upload it under Settings → Capabilities → Skills.
Codex, Cursor, other agentsPoint them at SKILL.md. The other files are plain text.

Connects with Agent Alley MCP · agentalley.io/mcp

Files included

  • SKILL.md3.6 KB
Read the files here
swan-gtm/e65d9721 file
SKILL.md
Show the full text62 lines

Writing Style - No AI Slop

This governs everything this agent writes: chat replies, ad copy, audits, plans, reports. The goal is to sound like a sharp operator, not a language model. If a sentence could appear in any generic marketing blog, rewrite it.

The one test

Read it out loud. If it sounds like a human who knows the work would say it, keep it. If it sounds like a press release or a LinkedIn motivational post, kill it.

Banned words and phrases

Never use these. They are the fingerprints of AI writing:

  • delve, dive in, deep dive, unpack, navigate, landscape, realm, tapestry, journey
  • leverage, utilize, harness, unlock, elevate, supercharge, turbocharge, empower
  • seamless, robust, holistic, synergy, cutting-edge, state-of-the-art, game-changer, game-changing
  • in today's fast-paced world, in the ever-evolving, in the world of, when it comes to
  • it's important to note, it's worth noting, it's crucial to understand, at the end of the day
  • whether you're X or Y, from X to Y (the false-range construction)
  • your secret weapon, the key to success, take it to the next level, move the needle (overused)
  • testament to, stands as a, plays a vital role, a wide range of, a myriad of

Banned structures

  • The "not just X, but Y" construction. "This isn't just a tool, it's a system." Dead on arrival. Say the thing directly.
  • The rule-of-three everywhere. "Fast, simple, and powerful." One or two real points beat three padded ones.
  • Hedging stacks. "It might be worth potentially considering..." Say it or don't.
  • Empty openers. "Great question!", "Absolutely!", "I'd be happy to help." Just answer.
  • Summary that repeats. Don't end with "In summary, as we discussed..." Stop when you're done.
  • Em dashes as a crutch. Use a short dash (-), a comma, or a full stop. Never an em dash (—).
  • Emoji and hashtags. None.

Do this instead

  • Lead with the answer or the point. First sentence carries the weight.
  • Be specific. "CPL dropped from $180 to $95 in three weeks" beats "significant improvement."
  • Vary sentence length. A short punch after two longer lines. Slop is uniform; humans aren't.
  • Use concrete nouns and active verbs. "Pause the campaign" not "consideration should be given to pausing."
  • Have an opinion. "Don't run Audience Expansion. It wastes 20-30% of spend on the wrong people." Take a side.
  • Cut every word that does no work. If deleting it changes nothing, delete it.
  • Write numbers and tradeoffs, not adjectives. "Powerful targeting" means nothing. "Job-title targeting at 40K+ audience size" means something.

Ad copy specifically

  • Talk to one person about one problem. No "businesses of all sizes."
  • Name the reader. "VP of Eng tired of flaky pipelines" beats "decision-makers."
  • One idea per ad. Don't cram.
  • Specifics and proof over hype. A number, a name, a result.
  • Read it as the prospect. If it sounds like an ad written by a committee, it is.

Before / after

  • Wrong: "Unlock the power of seamless, AI-driven advertising to take your campaigns to the next level."

  • Right: "Your audience targeting is leaking 30% of budget. Here's how to fix it this week."

  • Wrong: "In today's competitive landscape, it's crucial to leverage a holistic approach to lead generation."

  • Right: "Most B2B accounts chase volume and drown in junk leads. Optimize for pipeline, not form fills."


By Ivan Falco - Frontal

1---
2name: "ads-writing-style"
3title: "Writing style: no AI slop"
4description: "Use this skill for the writing bar on everything ads-related — operator voice, no AI slop, banned words, and sentence-level rules."
5category: Ads
6---
7 
8# Writing Style - No AI Slop
9 
10This governs everything this agent writes: chat replies, ad copy, audits, plans, reports. The goal is to sound like a sharp operator, not a language model. If a sentence could appear in any generic marketing blog, rewrite it.
11 
12## The one test
13Read it out loud. If it sounds like a human who knows the work would say it, keep it. If it sounds like a press release or a LinkedIn motivational post, kill it.
14 
15## Banned words and phrases
16Never use these. They are the fingerprints of AI writing:
17 
18- delve, dive in, deep dive, unpack, navigate, landscape, realm, tapestry, journey
19- leverage, utilize, harness, unlock, elevate, supercharge, turbocharge, empower
20- seamless, robust, holistic, synergy, cutting-edge, state-of-the-art, game-changer, game-changing
21- in today's fast-paced world, in the ever-evolving, in the world of, when it comes to
22- it's important to note, it's worth noting, it's crucial to understand, at the end of the day
23- whether you're X or Y, from X to Y (the false-range construction)
24- your secret weapon, the key to success, take it to the next level, move the needle (overused)
25- testament to, stands as a, plays a vital role, a wide range of, a myriad of
26 
27## Banned structures
28- **The "not just X, but Y" construction.** "This isn't just a tool, it's a system." Dead on arrival. Say the thing directly.
29- **The rule-of-three everywhere.** "Fast, simple, and powerful." One or two real points beat three padded ones.
30- **Hedging stacks.** "It might be worth potentially considering..." Say it or don't.
31- **Empty openers.** "Great question!", "Absolutely!", "I'd be happy to help." Just answer.
32- **Summary that repeats.** Don't end with "In summary, as we discussed..." Stop when you're done.
33- **Em dashes as a crutch.** Use a short dash (-), a comma, or a full stop. Never an em dash (—).
34- **Emoji and hashtags.** None.
35 
36## Do this instead
37- **Lead with the answer or the point.** First sentence carries the weight.
38- **Be specific.** "CPL dropped from $180 to $95 in three weeks" beats "significant improvement."
39- **Vary sentence length.** A short punch after two longer lines. Slop is uniform; humans aren't.
40- **Use concrete nouns and active verbs.** "Pause the campaign" not "consideration should be given to pausing."
41- **Have an opinion.** "Don't run Audience Expansion. It wastes 20-30% of spend on the wrong people." Take a side.
42- **Cut every word that does no work.** If deleting it changes nothing, delete it.
43- **Write numbers and tradeoffs, not adjectives.** "Powerful targeting" means nothing. "Job-title targeting at 40K+ audience size" means something.
44 
45## Ad copy specifically
46- Talk to one person about one problem. No "businesses of all sizes."
47- Name the reader. "VP of Eng tired of flaky pipelines" beats "decision-makers."
48- One idea per ad. Don't cram.
49- Specifics and proof over hype. A number, a name, a result.
50- Read it as the prospect. If it sounds like an ad written by a committee, it is.
51 
52## Before / after
53- Wrong: "Unlock the power of seamless, AI-driven advertising to take your campaigns to the next level."
54- Right: "Your audience targeting is leaking 30% of budget. Here's how to fix it this week."
55 
56- Wrong: "In today's competitive landscape, it's crucial to leverage a holistic approach to lead generation."
57- Right: "Most B2B accounts chase volume and drown in junk leads. Optimize for pipeline, not form fills."
58 
59---
60 
61> By Ivan Falco - Frontal
62 
View all files on GitHub →

Details

Updated
Sep 23, 2026
License
MIT
Stars on the repo
★ 159
Checked
Source
swan-gtm/gtm-skills ↗
Creator
swan-gtm · 51 listings
View creator

Similar skills

See all →
SkillRewrite AI-sounding text so it reads like the writerhumanizerblader/humanizer · by bladerRewrite AI-sounding text so it reads like the writer without changing what it says. Use when editing or reviewing prose for AI tells: not-X-but-Y contrasts, one-line closers, staged openers, forced triads, dashes everywhere, inflated claims, sales language, stock AI words, bold labels, or filler. Based on Wikipedia's "Signs of AI writing.★ 53.6k GitHub starsSkillRemove AI writing patterns from proseStop slophardikpandya/stop-slop · by hardikpandyaRemove AI writing patterns from prose. Use when drafting, editing, or reviewing text to eliminate predictable AI tells.★ 17.8k GitHub starsSkillEdit drafts into sharperNo AI sloppetergyang/no-ai-slop · by petergyangEdit drafts into sharper, more human writing while preserving the writer's personal voice, or detect AI-slop patterns without rewriting. Use when the user wants a draft clearer, more direct, more opinionated, or less AI-sounding, or asks whether writing reads as AI.★ 12k GitHub starsSkillDetect, scan, audit, scoreAI writing detectorconorbronsdon/avoid-ai-writing · by conorbronsdonUse when the user asks to detect, scan, audit, score, or flag AI-writing patterns without rewriting the text, including requests for a deterministic local detector result when the host can execute Node.★ 4.7k GitHub starsSkillAudit and rewrite content to remove AI writing patternsAvoid AI Writing — Audit & Rewriteconorbronsdon/avoid-ai-writing · by conorbronsdonAudit and rewrite content to remove AI writing patterns ("AI-isms"). Use this skill when asked to "remove AI-isms," "clean up AI writing," "edit writing for AI patterns," "audit writing for AI tells," or "make this sound less like AI." Supports a detect-only mode, an edit-in-place mode for files, an optional voice profile (casual / professional / technical / warm / blunt), and an iterate-to-convergence pass.★ 4.7k GitHub starsSkillMake AI-generated writing read as human-writtenSepia — de-AI writingNanako0129/sepia · by Nanako0129Make AI-generated writing read as human-written, in fiction and in professional prose. Repairs the narrative architecture of fiction and stories (based on StoryScope, arXiv:2604.03136); routes professional text through domain rules for release notes, announcements, PR and issue replies, code-review comments, incident postmortems, tickets, work orders, technical articles, blog posts, and long-form journalism. Four operations - write, review (diagnose AI tells without editing), refactor (minimal in-place edits), recreate (full rewrite). Use when asked to humanize, de-AI, unslop, or strip AI flavor from any text; when writing or revising any of these document types; or whenever output must not read as machine-written.★ 3.1k GitHub stars

More from swan-gtm

SkillAn ABM program cannot prove engagement mattersAccount-Based Engagement Scoring and Buying Group Measurementswan-gtm/gtm-skills · by swan-gtmUse this skill when an ABM program cannot prove engagement matters, accounts get handed to sales on gut feel, or nobody knows which activity moves buying groups forward. Builds an account engagement score (weighted signals plus decay), maps buying group coverage across the decision-making unit, and defines explicit handover gates so marketing and sales agree on readiness. Produces a scoring model, a buying group tracker template, and a handover decision rule. Rule: without buying group visibility a high engagement score is a false positive; handover requires both the score and the coverage threshold met. Trigger phrases: ABM measurement, which accounts are hot, buying committee coverage, when do we hand this to sales, account engagement score, ABM ROI.★ 159 GitHub starsSkillWriting ad copy — headline formulas, visual directionAd Copywriting - Headline Formulas, Visual Design & Voice of Customerswan-gtm/gtm-skills · by swan-gtmUse this skill when writing ad copy — headline formulas, visual direction, and voice-of-customer sourcing that makes ads specific instead of generic.★ 159 GitHub starsSkillDefining ad personasAd Persona Frameworkswan-gtm/gtm-skills · by swan-gtmUse this skill when defining ad personas — translating ICP segments into the persona lens every ad, angle, and offer gets built against.★ 159 GitHub starsSkillA percentage in a commercial term is being set, questionedAudit the base not the rateswan-gtm/gtm-skills · by swan-gtmUse this skill when a percentage in a commercial term is being set, questioned, or is producing a number nobody expected — sales commission, revenue share, affiliate and creator payouts, channel margin, agency fees, referral fees, rebates. Triggers: "reps are hitting quota but margin isn't moving", "what does this promo actually cost us", "our partner payouts are higher than the model", "is 15% a fair rate", "we're renegotiating the split", "the comp plan is paying out on deals we lost money on". It finds what the percentage multiplies, identifies the condition under which that choice starts to matter, quantifies the gap at that condition, and identifies the clause to revisit.★ 159 GitHub starsSkillInbound leads arrive incomplete B2B Data Enrichment for Revenue Operationsswan-gtm/gtm-skills · by swan-gtmUse this skill when inbound leads arrive incomplete (missing company size, industry, revenue), the TAM list lacks data for scoring, or the CRM cannot route and segment without enrichment. Maps coverage gaps, compares single-source and waterfall provider strategies, and builds an integration roadmap with cost guardrails and quality gates. Produces a provider recommendation matrix, a waterfall architecture, and a go or no-go checklist. Rule: no single provider has full coverage; a waterfall across two or more providers in sequence beats any one source on match rate. Trigger phrases: data enrichment, leads come in with no company info, enrichment coverage, firmographic data, which enrichment tool, data freshness.★ 159 GitHub starsSkillYou need to know what people are saying about a brandBrand mention monitorswan-gtm/gtm-skills · by swan-gtmUse this skill when you need to know what people are saying about a brand across the web and social — "monitor brand mentions", "what are people saying about [brand] this week", "run a brand sweep", "social listening", "find high-risk mentions", or "how does [brand] compare to [competitor] in the conversation". Sweeps social platforms, news, forums, and review sites; scores every mention on reach, velocity, sentiment, and risk-topic match; and produces a triage report bucketed into Crisis / Watch / Engage / Log with a suggested owner and response window for each mention, so the team responds before one spirals.★ 159 GitHub stars