Home / [048]
Design / writing-style skills for Claude Code / Codex developers that strip out the “AI taste”: free and open source to earn stars and an audience. Open gaps: non-web UI, specific frameworks, non-English writing styles, and turning taste into a linter
Open Trackable · rising open-source opportunity First seen 2026-11-01 · last updated 2026-11-02
- Status
- Open (the whole cluster is still growing, but it has split: the stretch backed by a platform keeps going, while the wave from dormant repos waking up is a spike; generic design skills are already taken)
- Why now, and how long the window lasts
- "AI slop" became a public topic this month. A platform (Perplexity) built one person's open-source skill in as a native feature and subsidized users. A skill is just a few markdown files that one person can build in a weekend. The window is measured in weeks: 10-25 is when the curve bent, and comparing star gains over 11-15 vs. 7 days will tell a spike from a trend.
- Confidence
- Medium. "The cluster is growing" is high-confidence. "Non-English writing styles are an open gap" is inferred and keyword searches may have missed something. Would be overturned by: at the 11-15 recheck, impeccable, hallmark and EGM each gain less than half their star count from the first 7 days over the 7-day window (meaning it was a spike); or Anthropic / OpenAI build design guidelines in as a native feature.
Evidence
Grouped by store. Each group starts with the charts that can be graphed (real curves, data dates below the chart), then the scanner's original notes and source links.
GitHub4 evidence · 18 products with curves
GitHub · created 2025-11-16 · 105,381 stars · 7 days +7,226
The design vocabulary for your coding agent. Steer with polish and bolder, review with critique and audit, catch AI slop before it ships.
Data 2026-08-27 → 2026-11-08 (74 days, 23 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-04-27 · 50,145 stars · 7 days +6,311
Anti-AI-slop design skill for Claude Code, Cursor, and Codex.
Data 2026-07-29 → 2026-11-08 (103 days, 17 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-09-13 · 13,694 stars · 7 days +6,169
Open-source skills for designing and reviewing software with AI agents
Data 2026-10-21 → 2026-11-08 (19 days, 17 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-07-29 · 14,830 stars · 7 days +1,730
Data 2026-10-09 → 2026-11-08 (31 days, 30 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-10-29 · 11,274 stars · 7 days +7,989
A Claude skill that turns vibecoded, AI-looking UI into a designed interface
Data 2026-10-31 → 2026-11-08 (9 days, 9 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-10-30 · 5,806 stars · 7 days +4,482
244 pixel art effects, every one a small piece of code. Plus a Claude Code skill to make your own.
Data 2026-10-31 → 2026-11-08 (9 days, 9 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-07-10 · 11,460 stars · 7 days +1,876
A collection of agent skills that help you build great interfaces.
Data 2026-07-30 → 2026-11-08 (102 days, 14 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-08-07 · 8,584 stars · 7 days +827
Rules for an AI coding agent to filter out generic AI-generated UI designs, text, and code.
Data 2026-10-09 → 2026-11-08 (31 days, 31 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
10 more products' curves
GitHub · created 2026-09-30 · 2,908 stars · 7 days +355
UI design skill for Claude Code, Codex and other AI agents: explore distinct directions, compare them side by side, refine against real screenshots. For AI agents
Data 2026-10-09 → 2026-11-08 (31 days, 31 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-08-20 · 6,762 stars · 7 days +2,066
一个基于 520 万字语料统计的去 AI 味 skill · An AI-tone removal skill grounded in a 5.2-million-character corpus study
Data 2026-10-09 → 2026-11-08 (31 days, 30 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-08-05 · 4,463 stars · 7 days +169
Make Chinese written by AI read like a specific person talking. A general-purpose writing and editing skill, ready to use out of the box.
Data 2026-10-09 → 2026-11-08 (31 days, 31 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-08-28 · 3,677 stars · 7 days +112
De-AI writing skill for any Agent Skills-compatible agent (77+ via the Skills CLI), with native plugins for Claude Code, Codex, Grok Build, and Antigravity. Nar
Data 2026-09-26 → 2026-11-08 (44 days, 32 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-09-30 · 2,896 stars · 7 days +122
An Agent Skill for Refining AI-Generated Japanese into Natural Japanese
Data 2026-10-01 → 2026-11-08 (39 days, 32 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-07-12 · 2,026 stars · 7 days +24
An Agent Skill for writing and revising work-related Japanese so it is easy to read and easy to understand.
Data 2026-10-09 → 2026-11-08 (31 days, 14 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-10-24 · 192 stars · 7 days +73
Membuat tulisan AI lebih luwes dan enak dibaca, bukan untuk mengelabui deteksi AI
Data 2026-10-25 → 2026-11-08 (15 days, 15 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-08-12 · 5,993 stars · 7 days +199
Opinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patterns
Data 2026-08-28 → 2026-11-08 (73 days, 32 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-09-02 · 126 stars · 7 days +2
ESLint rules for guarding AI slops in code.
Data 2026-10-09 → 2026-11-08 (31 days, 31 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
GitHub · created 2026-08-13 · 4,954 stars · 7 days +1,052
A builder, not just a researcher. Agent skills that turn top-grossing app patterns into native-quality mobile screens.
Data 2026-08-30 → 2026-11-08 (71 days, 32 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
- GitHub (full-site snapshot, verified 11-02): pbakaus/impeccable 99,169 stars (the report says 11-01 at 98,155), +10,818,10-25 from day 7 onward, still +1,000–2,000,11-02 per day after that, +1,014 in total. Nutlope/hallmark +9,804 over 44,695,7 days; its daily gain dropped from 2,800 at 10-27 to 861 at 11-02. google/elephant-goldfish-model +8,170,11-02 over 8,807,7 days and still +1,282.
- GitHub (verified 11-02): elayadesign/ai-design-skills gained +1,915 / +2,714 / +2,714 per day over 13,271,10-26~28, then 1,077 → 704 → 848 → 284 → 171. It was a spike and has already cooled. heydamir/unslop (created 10-29) +4,698,11-01 / +2,460 / +11-02 / +1,413, flagged as "new account, single repo surge." patch-gg/pixelfx (created 10-30) 2,446. jakubkrehel/skills 9,991. miqdadbadjuber/anti-slop 7,903 (+746 over 7 days, flat from last week). oil-oil/oil-ui 2,586.
- GitHub (writing style, verified 11-02): larashero3-dotcom/lieflat-less-ai-tone (Chinese) +659 over 4,735,7 days, +1,172 last week, cooling off. KKKKhazix/human-writing (Chinese) 4,308. Nanako0129/sepia (English) 3,570. Japanese nanaism/yomiyasu 2,793 (gaining +20–48 per day, steady). coji/natural-japanese 2,002 (no longer growing). Searching by humanizer / de-AI-ify / Korean, Spanish, German, French keywords, the only new repo with ≥100 stars is an Indonesian one, adhiprib/humanizer-indonesia (122 stars in 9 days). Korean, Spanish, German and French still have none.
- GitHub (taste turned into lint and non-web UI, verified 11-02): dmmulroy/anti-slop (Oxlint rules) +120 in 5,814,7 days. antfu/eslint-plugin-slop at 124 stars, no longer growing. Native mobile UI has only Appllama/appllama-skills 3,924 (+883 in 7 days, +151 last week, accelerating).
X1 pieces of evidence
- X (10-24, figures from the report, not verified): Perplexity's CEO posted "/impeccable is now a skill inside Computer…offering 5000 credits per user", getting 387 likes. impeccable accelerated again from the second day on.
Hacker News1 evidence items
- HN (verified 11-02): "Tells of a Slop UI" 2026-09-27 points 386. The report separately notes 10-22~30 had 8 titles mentioning slop.
For sale listings 1 pieces of evidence · 3 products with curves
TrustMRR · 30-day revenue $9,558 (+8%) · MRR $9,207 · 888 subscriptions
2025-11 → 2026-11 (13 pts). Light bars = current month, not yet complete
Data 2026-09-29 → 2026-11-06 (39 days, 39 observations). Single-day swings are big, so look at the trend in the monthly breakdown above.
WriteHybrid writehybrid.com
TrustMRR · 30-day revenue $4,908 (-30%) · MRR $5,111 · 241 subscriptions
2025-10 → 2026-05 (8 points)
Data 2026-09-29 → 2026-11-06 (39 days, 39 observations). Single-day swings are big, so look at the trend in the monthly breakdown above.
GenZWrite genzwrite.com
TrustMRR · 30-day revenue $3,823 (+65%) · MRR $3,053 · 613 subscriptions
2025-11 → 2026-09 (11 points)
Data 2026-09-29 → 2026-11-06 (39 days, 39 observations). Single-day swings are big, so look at the trend in the monthly breakdown above.
- Listed for sale (Acquire, 11-02, for comparison): consumer "rewrite it in plain language" tools are exiting. ag8qax9ofw (iOS app, TTM revenue $116,700, last month $8,500, asking $299,000); 8kg90xiqku (SaaS, TTM $48,500, last month $1,500, −62% year over year, asking $29,500; the seller admits it's declining). Of the 11 humanizers with revenue in indie, 7 are listed (Humanize AI $9,320, WriteHybrid $4,908 −30%, GenZWrite $3,823).
Already doing it
Generic design skills for the web UI: impeccable, hallmark, Google EGM, jakubkrehel/skills, unslop, oil-ui, pixelfx. Rules / lint: two anti-slop repos. Writing style: lieflat and human-writing (Chinese), sepia (English), yomiyasu and natural-japanese (Japanese), humanizer-indonesia (Indonesian). Native mobile UI: only Appllama.
The hardest step
The quality of the taste and rules has to be clearly better than impeccable on some specific medium. The second thing is monetization: the skill itself is free, and within a week dozens of copies will appear, so money can only come from the funnel or a paid library (see [049], [009]).
Who it's for
Developers with a sense of design taste who know a non-web framework (SwiftUI, Compose, Flutter, Qt, Unity, TUI); or writers whose native language is Korean, Spanish, German, or French. Open source and free, trading for stars and audience rather than direct revenue.
Should every product builder know this?
Yes. Users can already spot AI-generated interfaces and text at a glance, and every product gets judged that way.
Change log
2026-11-02
Backfill check: impeccable 98,155 → 99,169; hallmark 43,834 → 44,695; ai-design-skills 13,100 → 13,271, daily gain fell from 2,714 to 171, the spike is over; EGM 7,525 → 8,807; unslop 3,285 → 4,698. The whole cluster is still growing, but it has split: impeccable, which has platform backing, and Google's own EGM still gain 1,000+ per day, while the dormant-library wake-up wave has receded. One non-English style was added (Indonesian); Korean, Spanish, German, and French remain open.
2026-11-01
First recorded (backfill: from «New GitHub Opportunities (Batch 1)» §1.1, «Stringers» §3.1 / §3.2, «Exit Listings (Batch 1)» §2.6; data as of 11-01 / 11-02).
Original: Scanner opportunity library #048 (updated daily, 06:00); this page was generated on the 2026-11-08 edition.