SkillMesh – AI-powered Agent Skills Marketplace

Overview

SkillMesh is a chat-first, AI-powered platform where users publish, discover, and exchange Agent Skills. Users submit requests for specific skills, and AI automatically matches them with the most relevant skills and opens a conversation. Skills can be activated directly in the chat, with built-in TNK incentives to reward skill authors and platform usage.

Problem it solves

Agents are capable, but often lack the context or procedural knowledge needed to deliver results reliably. SkillMesh allows anyone to package knowledge, workflows, or scripts into reusable Agent Skills that can be discovered, activated, and used in conversations.

How it works

1. Publish skills – Users create a profile and upload skills (instructions, scripts, workflows) in a standard format (SKILL.md).
2. Request a skill – Another user posts a request describing what they need.
3. AI match – AI compares the request to all skills and selects the best matches.
4. Auto conversation – AI opens a chat with matched users/skills, showing skill metadata and instructions.
5. TNK integration – Activating a skill costs TNK. Part of the TNK goes to the skill author, part to the platform.

MVP Features

• ✅ Intercom-style inbox & chat
• ✅ User profiles with skills offered/needed
• ✅ AI-powered skill matching
• ✅ Skill activation & TNK tracking
• ✅ TNK Earnings dashboard

TNK Economics

• Skill activation is charged in TNK
• Author receives a portion of activation cost
• Platform receives a portion for maintenance & operations

Trac Integration

Trac Address: bc1pjax699xyckyk8srqc02z2nc2za0zejq04malp2erxgq2rrjwvj9qsre5vu

All TNK payments and earnings are tracked on the Trac network.

Project Structure

SkillMesh/
├── README.md (this file)
├── SKILL.md (example skill format)
├── docs/
│   ├── ARCHITECTURE.md
│   └── TNK_INTEGRATION.md
├── src/
│   ├── components/ (UI components)
│   ├── services/ (API & Trac integration)
│   └── hooks/ (React hooks)
└── public/

Running the Demo

See the demo video: [Link to Loom/Screen recording]

The demo shows:

1. User uploads a skill
2. Another user requests a skill
3. AI matches & opens a chat
4. Skill metadata is displayed
5. TNK activation logged

Getting Started

[Coming soon - development setup instructions]

Contributing

Submit a pull request with your improvements!

---

Built for the Trac Network Competition