35,581
Top Skill Downloads
Capability Evolver #1
1.5M+
Total ClawHub Downloads
across 3,286 skills
133K
Top Weekly Installs
Vercel Web Design
$0.001
Lowest Per-Call Price
→ $1.00 highest
1 Top Skills by Real Downloads
ClawHub verified install counts — the only marketplace publishing exact numbers
ClawHub — Top 20 by Downloads
1.5M+ total platform downloads| # | Skill | Downloads | Consumption Bar | Category | ★ |
|---|---|---|---|---|---|
| 01 | Capability Evolver | 35,581 | AI / ML | 33 | |
| 02 | Wacli | 16,415 | Utility | 37 | |
| 03 | ByteRover | 16,004 | Utility | 36 | |
| 04 | Self-Improving Agent | 15,962 | AI / ML | 132 | |
| 05 | ATXP | 14,453 | Utility | — | |
| 06 | GOG (Google Workspace) | 14,313 | Productivity | 48 | |
| 07 | Agent Browser | 11,836 | Web | 43 | |
| 08 | Summarize | 10,956 | Productivity | — | |
| 09 | GitHub | 10,611 | Dev Tools | — | |
| 10 | SonosCLI | 10,304 | Media | — | |
| 11 | Weather | 9,002 | Productivity | — | |
| 12 | Humanize AI Text | 8,771 | Writing | — | |
| 13 | Tavily Web Search | 8,142 | Data & Research | — | |
| 14 | Free Ride | 7,927 | Utility | — | |
| 15 | Bird | 7,767 | Utility | — | |
| 16 | Find Skills | 7,077 | Meta | — | |
| 17 | Proactive Agent | 7,010 | AI / ML | — | |
| 18 | Auto-Updater Skill | 6,601 | Meta | — | |
| 19 | Obsidian | 5,791 | Productivity | — | |
| 20 | Nano Banana Pro | 5,704 | Utility | — |
Pattern: AI/ML agent skills (self-optimization, memory) dominate the top 5. Utility tools cluster at #2–5.
Productivity (Google, Summarize, Obsidian) is the broadest category. Dev tools underperform on raw downloads but
have highest per-user engagement. Self-Improving Agent has 132★ — 4x more stars-per-download than #1.
2 Claude Code & Cross-Platform Skills
Weekly installs across Claude Code, Codex, and compatible agents
Top Skills by Weekly Installs
Cross-platform| # | Skill | Weekly Installs | Consumption | Type | Source |
|---|---|---|---|---|---|
| 01 | Vercel Web Design Guidelines | 133,000 | Design | Community | |
| 02 | Webapp Testing | 117,000 | Testing | Official | |
| 03 | Firecrawl | 110,000+ | Data & Research | Community | |
| 04 | Everything Claude Code (ECC) | — | Collection | 163K ★ | |
| 05 | Superpowers | — | Collection | 40.9K ★ |
Anthropic Official Skills
17 skills · 125.5K ★| Skill | Category | Note |
|---|---|---|
| Webapp Testing | Testing | 117K weekly — most used official skill |
| Word Document | Docs | Powers Claude.ai document features |
| Excel Document | Docs | Powers Claude.ai spreadsheet features |
| PDF Document | Docs | Powers Claude.ai PDF features |
| MCP Server Creator | Dev Tools | MCP server generation |
| Skill Creator | Meta | SKILL.md scaffolding |
Agensi Marketplace (exact installs)
Precise counts| # | Skill | Installs |
|---|---|---|
| 1 | code-reviewer | 116 |
| 2 | git-commit-writer | 65 |
| 3 | readme-generator | 49 |
| 4 | pr-description-writer | 36 |
| 5 | temporal-reasoning-sleuth | 32 |
| 6 | env-doctor | 30 |
| 7 | changelog-generator | 27 |
3 Real Pricing Data
Verified per-call costs and subscription tiers from active MCP services
Per-Call Micropayments
x402 / Gateway / API| # | Service | $/call | Scale |
|---|---|---|---|
| 01 | Weather endpoint | $0.001 | |
| 02 | Firecrawl (basic scrape) | $0.005 | |
| 03 | Tavily search (pay-go) | $0.008 | |
| 04 | Ref doc search | $0.009 | |
| 05 | Exa / generic web search | $0.01 | |
| 06 | Pain-point analysis | $0.02 | |
| 07 | Code analysis | $0.05 | |
| 08 | Firecrawl (crawl+extract) | $0.037 | |
| 09 | Research paper | $0.10 | |
| 10 | Dataset query | $0.25 | |
| 11 | Validator report (Solana) | $1.00 |
Tavily �� Full Pricing Tiers
Most used paid MCP| Plan | Credits/mo | $/mo | $/credit |
|---|---|---|---|
| Free | 1,000 | $0 | — |
| Project | 4,000 | $30 | $0.0075 |
| Bootstrap | 15,000 | $100 | $0.0067 |
| Startup | 38,000 | $220 | $0.0058 |
| Growth | 100,000 | $500 | $0.005 |
1 credit = basic search · 2 credits = advanced · 15–250 credits = pro research
Subscription & Credits Models
Verified tiers| Service | Free Tier | Hobby/Pro | Growth | Notes |
|---|---|---|---|---|
| Firecrawl | 500 credits | $16/mo (3K credits) | $333/mo (500K) | +$9/1K overage |
| Composio | 20K calls/mo | $29/mo (200K calls) | $499/mo (RBAC) | SaaS integrations |
| Ref (docs search) | 200 credits ∞ | $9/mo (1K credits) | $9 /1K block | Never-expire free tier |
| 21st.dev Magic | 10 credits | $16/mo | $32/mo | UI component generation |
| Taskade MCP | Yes | $6/mo | $40/mo | Workspace productivity |
| PatSnap | 10K credits | Pay-as-you-go | Enterprise | Patent/R&D data |
| OpenClaw Cloud | Self-host free | $29.50 (first mo) | $59/mo | Platform hosting |
95% of actual usage still runs on the API-key model — user pastes a key, agent calls tools, human pays the provider bill.
Credits-based models are the emerging standard for indie MCP services. x402 micropayments exist but remain experimental.
4 Payment Protocol Stack
Layered, not competing — each protocol solves a different part of the transaction
Agentic Commerce Protocol
Merchant checkout — bridges agents into card-based e-commerce
Agent Payments Protocol
Authorization mandates — cryptographic proof of user approval
x402 — HTTP Payment Required
Per-call settlement via USDC — no accounts, no API keys needed
L402 — Lightning Auth
Per-request via Bitcoin Lightning — zero-commitment, sub-cent fees
Non-negotiable: Spending policies must be enforced in code, not in prompts.
An LLM can be social-engineered. A hard cap in the wallet layer cannot.
5 Market Segments & Strategy
Where the market stands, who's making money, and what matters next
Segment Matrix
6 segments| Segment | Model | Key Players | Maturity | Revenue |
|---|---|---|---|---|
| Free Skills | GitHub distribution | ClawHub, SkillHub, Anthropic | Mature | 0% |
| Curated / Review | Community + AI scoring | SkillJury, SkillHub | Growing | 0% |
| Paid MCP Services | API key (95%) | Tavily, Firecrawl, Composio | Early-mid | <5% |
| Micropayments | x402 + gateway | xpay, mcp-billing-gateway | Early | Experimental |
| Enterprise Gateways | Policy-controlled | AWS, Azure, TrueFoundry | Early | Enterprise |
| Agentic Wallets | Autonomous | Coinbase | Very early | 18–24 mo |
◆
Discovery is the moat
Skills commoditize fast — they're text files on GitHub. The moat sits in who controls how agents find them: registries, gateways, ranking.
⏲
Payment rails still settling
Path of least resistance today: API keys + gateway metering. AP2/x402 for direct agent commerce is 18–24 months out for enterprise.
◎
Descriptions win selection
Tool selection = ~200ms from ~100 tokens. The team that writes the cleanest SKILL.md description wins disproportionate selection share.