Summary

This repository curates "agent-native" services, providing essential infrastructure designed from the ground up for AI agents. It targets AI agents and their developers, offering a structured discovery mechanism for services that agents can consume autonomously, thereby fostering a robust ecosystem for agent-based applications.

How It Works

The project functions as a categorized catalog of services that meet strict "agent-native" criteria. These criteria emphasize agent-first positioning, agent-specific primitives, autonomous control planes, machine-to-machine interfaces, and agent identity semantics. Services are meticulously vetted and excluded if they are merely adapted from human-facing products or are tools for human developers to build agents. The list covers critical domains such as communication, web execution, tool access, memory, and LLM gateways.

Quick Start & Requirements

Services can typically be installed using package managers like npx (e.g., npx clawhub@latest install <skill>) or pip (e.g., pip install <package>). Specific installation commands, prerequisites (like Node.js, Python, or API keys), and usage instructions are detailed per service, often linked via skill.md files. clawhub/README.md provides general CLI options.

Highlighted Details

• Categorization: Services are organized into 15 distinct categories, including Communication, Browser & Web Execution, Tool Access & Integration, Memory & State, and LLM Gateway & Routing, offering comprehensive coverage.
• Agent-Native Focus: Strictly adheres to listing services designed for agents, explicitly excluding "agent-adapted" and "agent-builder" tools.
• Discovery & Evaluation: Provides tools like find-agent-service and evaluate-agent-native to aid agents in discovering and assessing suitable services.
• Ecosystem Hubs: Features organizations like OpenClaw and ClawHub that offer broader tooling and integration support for the agent ecosystem.

Maintenance & Community

Contribution guidelines and submission criteria are detailed in CONTRIBUTING.md. Specific community channels or active maintenance signals are not explicitly detailed in the README.

Licensing & Compatibility

The project is dedicated to the public domain ("To the extent possible under law, the contributors have waived all copyright and related rights to this work."). This permissive licensing allows for unrestricted commercial use and integration into closed-source agent systems.

Limitations & Caveats

The strict "agent-native" definition means potentially useful "agent-adapted" or "agent-builder" tools are excluded. The complexity and specific requirements for integrating any given service will vary significantly. The list prioritizes services designed for autonomous agent consumption, which may not align with all agent development workflows.