Skillforge
A commerce and verification layer for reusable coding-agent skills, with versioned packages, explicit permissions, bounded test evidence and author payouts.
Reusable coding-agent skills already have directories, marketplaces and package-based distribution. The supplied research explicitly rejects the original first-mover premise and confirms multiple live competitors plus an official marketplace system. The remaining gap is narrower: paid skills, revenue sharing and trustworthy evidence about what a particular version did under declared tests.
Skillforge lets an author publish a versioned package with ownership, license, dependency, permission and support declarations. A buyer can inspect automated test evidence, reviewer findings, compatibility claims and change history before installing. Purchase, entitlement, installation approval, runtime permission and execution remain separate.
A passed test is not universal quality, security certification, legal compliance or a guarantee that a skill will behave safely in another repository. Popularity and usage telemetry are not correctness. Dependencies can change, packages can be malicious and platform owners can add payments quickly. The product must compete on commerce governance and bounded evidence rather than directory size.
The first release should support one package format, one paid entitlement model and manual publisher review. It should not execute untrusted packages during listing, collect repository content by default, hide fees, certify security from a shallow scan or present self-improvement claims without explicit user control and versioned evidence.
Independent skill author or small developer team seeking paid distribution, and professional users buying supported reusable coding workflows
Rapid ecosystem growth and official marketplace support create a strong but high-risk window.
An open package format and available commerce infrastructure explain why a paid layer is feasible now.
The record has no cross-reference mentions, two inbound connections and no direct connections.
The supplied research confirms a rapidly active skills ecosystem and leaves a narrower monetization plus bounded-verification gap.
Multiple marketplaces already exist, the platform owner can add payments, buyer scale and budget are incomplete, quality review adds marginal cost and no structural incumbent copying cost is established.
Discussion
No comments yet — be the first to weigh in.
