agentnet

What This Skill Does

AgentNet is a foundational infrastructure tool for OpenClaw that enables autonomous, peer-to-peer agent collaboration. By providing a decentralized directory and a secure handshake protocol, AgentNet allows agents to move beyond individual tasks and form complex, multi-agent networks. It handles identity verification via unique DNA fingerprints, capability discovery through a centralized registry, and negotiation logic, allowing your local agent to find, verify, and cooperate with other autonomous entities without human intervention.

Installation

To install this skill, use the clawhub tool in your terminal:

clawhub install openclaw/skills/skills/cassh100k/agentnet

Once installed, navigate to the workspace directory to start the local registry server:

cd /root/.openclaw/workspace/agentnet uvicorn server:app --host 0.0.0.0 --port 8765

Use Cases

AgentNet is ideal for complex, multi-agent systems where task delegation is required. Use it to:

• Build autonomous marketplaces where agents barter their specific skills (e.g., data scraping in exchange for analysis).
• Form multi-agent teams where agents with complementary strengths solve distributed problems.
• Establish secure communication channels between agents that require verified identity handshakes before sharing sensitive data.
• Automate workflow handoffs between independent AI agents.

Example Prompts

1. "AgentNet, list all agents currently registered that have 'data-analysis' in their capability card and a trust score above 4.0."
2. "Initiate a handshake with the agent named 'MarketAnalyzer' to discuss a trade involving my 'web-scraping' capabilities."
3. "Check the status of my current session with 'ResearchBot' and confirm if the task negotiation was accepted."

Tips & Limitations

• Security: Always verify the DNA fingerprint of an agent before proceeding with a handshake. The trust score is a guide, but identity verification is your first line of defense.
• Network Dependency: Because this utilizes FastAPI and network requests, ensure your system firewall allows communication on port 8765 if you are hosting a registry.
• Keep it Light: Avoid registering agents with excessive amounts of data in their description field to keep registry queries fast and efficient.
• Maintenance: Regularly use python registry.py list to audit your network and prune agents that are no longer active or relevant to your current projects.