Skip to content
gentic.news — AI News Intelligence Platform
Connecting to the Living Graph…

Listen to today's AI briefing

Daily podcast — 5 min, AI-narrated summary of top stories

A stark bar chart showing extreme market concentration among top agent skill sellers, with a single dominant bar…
AI ResearchScore: 75

Claude Code Skill Market Gini Hits 0.96

Claude Code skill adoption is a winner-take-all market (Gini 0.96). Top 1% hold 77.4% of installs; median skill gets 56. Publish niche, bundle skills, or target the fast-growing interrogation category (grill-me: 787K installs).

·1d ago·5 min read··21 views·AI-Generated·Report error
Share:
Source: dev.tovia devto_claudecode, hn_claude_codeWidely Reported
How do I get my Claude Code skill adopted when 77% of installs go to the top 1%?

The agent skill market has a Gini coefficient of 0.96. The top 1% of skills hold 77.4% of all installs, and the median skill has just 56 installs. To compete, publish niche skills with clear documentation, or bundle multiple skills like ByteDance's Lark suite.

TL;DR

The agent skill market is brutally unequal: 198 skills hold half of all 158M installs. Publish niche, bundled, or interrogation-style skills to win.

Key Takeaways

  • Claude Code skill adoption is a winner-take-all market (Gini 0.96).
  • Top 1% hold 77.4% of installs; median skill gets 56.
  • Publish niche, bundle skills, or target the fast-growing interrogation category (grill-me: 787K installs).

What Changed — The Agent Economy Census, August 2026

On 2026-08-09, Skillselion dumped its entire production catalog: 79,827 public listings covering agent skills, MCP servers, plugins, and marketplaces — 157,833,061 cumulative installs. They then ran the Gini coefficient (the economic inequality metric) across the 58,751 listings that are agent skills.

The result: 0.96. That's not a typo. South Africa's income Gini is 0.63. Brazil's is 0.52. A 0.96 is closer to a raffle than a market.

For Claude Code users, this isn't abstract. The skills you install — or publish — live in this distribution.

What It Means For You — The Numbers That Matter

Here's what 0.96 looks like from the inside:

  • Half of all 158M installs go to just 198 skills — that's 0.34% of the catalog.
  • The median skill has 56 installs. Not 5,600. Fifty-six.
  • 59.3% have fewer than 100 installs. 9.9% have zero.
  • The top 1% hold 77.4% of every install.

What does this mean for your daily Claude Code workflow? When you install a skill, you're almost certainly pulling from that top 1%. When you publish one, you're entering a lottery where the median outcome is effectively zero.

But there's a second, more actionable layer: the fastest-growing category in the entire census is interrogation skills — tools that argue with you and grill your code. The cluster of 88 skills reached 2,425,925 installs and is growing at +160%, the #1 rising search in the platform.

Top interrogation skills:

  • grill-me: 787,155 installs (mattpocock/skills)
  • grill-with-docs: 669,499 installs
  • triage: 534,688 installs
  • grilling: 380,693 installs

Matt Pocock — an independent TypeScript educator — sits at #3 overall, above Vercel Labs, Anthropic, Google, and AWS combined. He's also the author of the top interrogation skills.

Try It Now — How to Compete in a 0.96 Market

From AI to Financial Intelligence: How to Connect Claude to Real Market ...

1. Install an interrogation skill today.

If you're not using grill-me or a similar skill, you're missing the fastest-growing workflow in the ecosystem. It's the #1 rising search for a reason: developers want their Claude Code agent to challenge assumptions, not just execute.

# From mattpocock/skills repo
claude --add-skill grill-me

2. If you publish skills, bundle them.

ByteDance's Lark/Feishu suite holds 15.4% of all installs — but they're installed as bundles. One user action installs several skills at once, and each gets counted separately. This is the single biggest lever in the census for boosting install counts.

3. Target a niche, not a generalist audience.

The median skill has 56 installs. Skills that automate work reporting — standups, meeting notes, status reports — collectively hit 1,889,988 installs across 111 skills. Meeting summary, inbox triage, and resume tailoring are proven categories.

4. Don't read your numbers as a verdict.

These are cumulative all-time totals. Anything published this week starts from zero against a field that's been accumulating for months. The median skill has 56 installs, so if you're below that, you're not failing — you're average.

5. Consider charging.

Of 79,827 listings, only 55 carry a price (0.07%). If you build something genuinely useful, you'd be one of 55. The market hasn't built a paid tier yet — that's an opportunity, not a barrier.

The Bottom Line

The agent skill economy is brutally unequal. But the census also shows that a single independent developer with TypeScript education skills can outrank Google and OpenAI. The market rewards what developers actually install, not what companies fund.

If you want to win: build interrogation tools, bundle your skills, target the work-reporting niche, and don't expect overnight adoption. The median is 56 installs — beat that and you're in the top 40%.

Full methodology and raw CSV available at Skillselion's Agent Economy Census.


Source: dev.to

[Updated 12 Aug via devto_claudecode]

Design skills are now surging alongside interrogation tools. On 2026-08-12, Skillselion reported that four of the seven fastest-rising Claude Code skill searches are design-focused (Ui Ux Pro Max, Frontend Design, Impeccable, Design Taste Frontend), collectively holding 1,659,726 installs and growing +20% to +70%. Anthropic's own Frontend Design leads with 767,871 installs, explicitly aimed at avoiding "templated defaults." Independent leonxlnx's Design Taste Frontend (350,625) and its sibling high-end-visual-design (265,273) show a family of taste skills gaining traction. [per Skillselion]

[Updated 13 Aug via devto_claudecode]

The taste-skill trend extends beyond the top four: leonxlnx's family includes high-end-visual-design (265,273 installs) and stitch-design-taste (223,682), both fetched from the live catalog on 2026-08-12. Meanwhile, non-design risers signal a restraint theme: Grill Me now at 832,797 installs, Caveman (429,065) targeting a 75% token cut, and Ponytail (33,353) enforcing minimal changes. [per Skillselion]


Sources cited in this article

  1. Skillselion
Source: gentic.news · · author= · citation.json

AI-assisted reporting. Generated by gentic.news from 1 verified source, fact-checked against the Living Graph of 4,300+ entities. Edited by Ala SMITH.

Following this story?

Get a weekly digest with AI predictions, trends, and analysis — free.

AI Analysis

Claude Code users should change their workflow in three concrete ways based on this census. First, immediately install a grill/interrogation skill (grill-me has 787K installs and is the fastest-growing category at +160%). This isn't a fad — it's the #1 rising search on the platform, meaning developers are actively seeking agents that challenge their code rather than just executing commands. Add it to your CLAUDE.md workflow: `Always run grill-me before finalizing any PR.` Second, if you publish skills or MCP servers, adopt ByteDance's bundling strategy. The census shows bundled skills inflate install counts dramatically — Lark/Feishu holds 15.4% of all installs through this mechanism. Package related skills together so one install action counts multiple times. Also target the work-reporting cluster (standups, meeting notes, status reports = 1.89M installs) — this is proven demand with less competition than generalist tools. Third, recalibrate your expectations around adoption. The median skill has 56 installs. If your published skill has a few hundred installs, you're in the top 40% — that's success, not failure. The market is a raffle (Gini 0.96), so don't judge your work by install counts. Focus on the 0.07% of listings that charge money — if you build something truly valuable, you'd be one of 55 paid skills in a market of 79,827. That's an opening, not a ceiling.
This story is part of
Hugging Face Becomes the Neutral Ground Where Google and Anthropic's Agent Protocol War Converges
As Claude Code's MCP dominance threatens Google Cloud, Hugging Face's unique position as partner to both players creates an unexpected convergence zone
Compare side-by-side
Claude Code vs Interrogation skills
Enjoyed this article?
Share:

AI Toolslive

Five one-click lenses on this article. Cached for 24h.

Pick a tool above to generate an instant lens on this article.

Related Articles

From the lab

The framework underneath this story

Every article on this site sits on top of one engine and one framework — both built by the lab.

More in AI Research

View all