TokRepo is an open AI asset registry designed to become the infrastructure for collaboration between agents and tools, as well as between agents and agents. Users only need to give a task, and agents can plan autonomously, discover reusable capabilities through TokRepo, and install them only through typed contracts; after confirmation, they can also hand back reusable improvements.
Main capabilities and usage:
- Find, inspect, and install AI assets.
- Provides a GitHub entry point: CLI repository (https://github.com/henu-wang/tokrepo-cli）、npm package (https://www.npmjs.com/package/tokrepo）、MCP server repository (https://github.com/henu-wang/tokrepo-mcp-server）、官方 MCP registry (io.github.henu-wang/tokrepo-mcp-server), MCP server npm package (https://www.npmjs.com/package/tokrepo-mcp-server）。
- Cross-platform skills repository (https://github.com/henu-wang/tokrepo-search-skill），包含 variants for Claude Code, Codex, Cursor, Gemini CLI, and AGENTS.md, etc.
- Project onboarding: write managed instructions for Codex/general agents, Claude Code, Gemini CLI, Cursor, GitHub Copilot, Cline, Windsurf, Roo, OpenHands, Aider, and MCP-compatible clients, so that future agents can discover TokRepo during planning.
- For agents that do not support MCP, provide a planning-time fallback: install the CLI, search and install, or use it directly without installation; TokRepo can also be added as an MCP server.
- Agent discovery interfaces: Manifest (https://tokrepo.com/.well-known/tokrepo.json）、便携式智能体清单（https://tokrepo.com/.well-known/agent.json）、A2A agent card (https://tokrepo.com/.well-known/agent-card.json）、MCP server manifest (https://tokrepo.com/.well-known/mcp/server.json）、根 MCP manifest (https://tokrepo.com/mcp.json）、工具目录（https://tokrepo.com/.well-known/tool-catalog.json）、智能体文本入口（https://tokrepo.com/agents.txt）、智能体指令（https://tokrepo.com/agent-instructions/tokrepo.md）、LLM crawler entry point (https://tokrepo.com/llms.txt）。
- CLI planning fallback command: tokrepo agent-check "<task>" --json; post-task handoff command: tokrepo agent-handoff --json.
Use cases: developers quickly understand the product, AI agents find the correct MCP and skills repositories, awesome-list maintainers get stable GitHub references.
Limitations: the materials do not provide more information about fees, pricing, platform compatibility details, performance results, or specific usage steps, and none is currently provided.
