Capafy has launched as an AI skills marketplace that lets creators publish proprietary skills and earn per-use payments across Claude Code, Codex, and OpenClaw. Unlike existing open-source skill registries where anyone can fork and redistribute published skills, Capafy executes skill logic server-side: buyers receive outputs but never see the files, code, or reasoning behind them, according to TestingCatalog.
The platform targets a structural gap in open-source agent ecosystems. Creators with valuable domain expertise have had little incentive to publish their best work publicly, because the moment a skill goes into an open registry, anyone can copy the methodology and redistribute it at zero cost. Capafy inverts that model by treating skill logic as intellectual property that stays on its servers.
How It Works
Creators set their own pricing, and every skill execution triggers a payment to the creator. On the user side, skills integrate with Claude Code, Codex, and OpenClaw in one click with no installation required, per TestingCatalog.
The Capafy marketplace is live with listings spanning video ad creation, social media copywriting, TikTok script generation, and content optimization. The initial creator roster includes specialists in e-commerce video production, social media hooks, and content strategy, each packaging domain-specific workflows that outperform generic agent defaults on their specific tasks.
A Crowded but Structurally Different Market
The agent skills marketplace has expanded rapidly, growing from a handful of open registries in late 2025 to more than eight active platforms by mid-2026, including Agensi, SkillsMP, ClawHub, and skills.sh, according to TestingCatalog. All of those platforms operate on an open-source assumption: skills are public, forkable, and free or purchased once downloaded.
Capafy’s closed-source model is a bet that the most valuable agent skills will come from domain experts who need protection for their intellectual property before they will publish at all. The platform positions itself around what it calls the “invisible 20%” of value creation: while generic agents handle baseline tasks, domain-specific expertise and calibrated methodologies produce measurably better output in specialized contexts.
The Creator Economy Question
The launch raises a question that the broader agent ecosystem has not yet answered: who captures the value when agents get good at specific tasks? Open-source registries optimize for distribution and community contribution. Capafy optimizes for creator compensation, betting that the same dynamic that drove Gumroad, Teachable, and Patreon to success in the content creator economy will apply to AI skill creators.
Creator accounts and skill listings are open globally as of launch.