AI agent networking, explained.
Guides and field notes on connecting AI agents over a real network: direct peer-to-peer communication, NAT traversal without a VPN, encrypted tunnels, per-peer trust models, and the Pilot Protocol app store — with working pilotctl commands in many posts.
Start with agent discovery over P2P networks, peer-to-peer agent communication without a server, how Pilot Protocol works under the hood, or an introduction to the app store.
Verify an Install Script Before Running It: What the shell.online Installer Checks, and What It Can't, After Tensorlake 0.5.144
How to verify an install script before running it, using the shell.online installer after the Tensorlake npm compromise: what the checksum proves and doesn't.
Terminal Sharing Security When Infostealers Go After Coding Agents: What shell.online Leaves on Disk, and What to Revoke
Terminal sharing security after Gen Digital's infostealer report: the two files shell.online keeps on a host, what each unlocks, and the cleanup order.
Keep a Terminal Session Alive After a Disconnect: What a Long Agent Run on shell.online Survives
How to keep a terminal session alive after a disconnect: what shell.online does when the phone, the network or the host drops during a long agent run.
Share a Dev Server's Terminal With a Teammate, Not Its Port: Restarting on Go 1.26.9 After Two net/http Fixes
Go 1.26.9 patches two net/http bugs. Share the dev server's terminal with a teammate in the browser while the server itself stays on localhost.
A Persistent Terminal Sharing Link That Survives a Reboot: shell.online --persistent After Ubuntu's October Kernel Update
Keep a persistent terminal sharing link through a reboot: what shell.online --persistent restores, what it does not, and how to guard the state file.
Two People, One Collaborative Terminal Session in the Browser: How shell.online Decides Who Types During a GhostAction Cleanup
How a collaborative terminal session in the browser decides who types: shell.online's 1.8-second input lease, read-only links and rotation.
A Remote Router Shell Without Port Forwarding or an Exposed SSH Port, After MikroTrick
A remote router shell without port forwarding: shell.online connects out from the device, so no SSH port waits on the WAN for the next MikroTrick.
A Terminal Sharing Link That Expires at a Set Time: What shell.online --auto-close Actually Stops
A terminal sharing link that expires on a deadline: how shell.online --auto-close stops an unattended agent, and what it does not cover.
Share a Windows Terminal Session in the Browser: Codex on ConPTY, Without OpenSSH or WSL
Share a Windows terminal session in the browser with shell.online: Codex on native ConPTY, no OpenSSH server or WSL, and the Windows-only limits.
A Browser Terminal for a Docker Container That Keeps One Link Across Restarts, After the Unsloth Studio RCE
A browser terminal for a Docker container with shell.online: one encrypted link across restarts, no SSH keys inside, and why the Unsloth RCE makes that matter.
See What a Coding Agent Is Doing Without Opening Its Terminal: Owner-Only Session Summaries in shell.online, After PixelLeak
See what a coding agent is doing from the session list: new shell.online summaries are sealed to the owner and unreadable by the service.
End-to-End Encrypted Terminal Sharing Has a Second Secret on the Host: shell.online v0.24.1 Takes the Session Password Out of Process Environments
What end-to-end encrypted terminal sharing hides from the relay, and why shell.online v0.24.1 moved the session password out of process environments.
Let a Coding Agent Share Its Own Terminal Session in the Browser: the shell.online Skill and shell --json
How the shell.online agent skill tells Claude Code or Codex to share a terminal session in the browser and hand you the link, read-only when it should be.
Watch a Coding Agent Remotely Without Shrinking Its Terminal: shell.online v0.23.2 Gives the Host Its Grid and Each Viewer Its Own Layout
How to watch a coding agent remotely from a phone now that shell.online v0.23.2 lets the host own its terminal grid and each browser viewer reflow its own view.
Self-Hosted Terminal Sharing: Run the shell.online Relay on One Docker Host Instead of Cloudflare
Cloudflare had a Durable Objects incident on 25 September. Here is self-hosted terminal sharing with the standalone shell.online relay, step by step.
Start a Terminal Session From the Browser on a Mac That Sat Idle for Days: the launchd Priority Fix in shell.online v0.23.3
Why browser-started sessions on an idle Mac lagged for seconds, and what v0.23.3 changes when you start a terminal session from the browser.
When the Egress Proxy Rotates Its Credentials
Inside Meta Muse the proxy credentials rotate every few minutes: long-lived processes start getting 407s while their old tunnels stay up. How to detect it, and a relay that fixes it.
Watch Claude Code From Your Phone as a Conversation, Not an 80-Column Grid
How to watch Claude Code from your phone with the shell.online chat renderer, what it reads from the screen, and where the transcript is kept.
Getting a Pilot Node Online From a Locked-Down Agent VM
How a Pilot Protocol node registered from inside Meta Muse's sandbox: no UDP, poisoned DNS, proxy-only egress, six dead ends, and the SNI router that worked.
Share a Terminal Session Password With a Teammate Without the Service Seeing It: shell.online v0.23.1 Password Requests
shell.online v0.23.1 lets a teammate ask a session owner for its password; the copy is sealed in the owner's browser and the service never holds it.
Which Agent Session Needs You? shell.online v0.22 Adds a Passive Pulse to Monitor a Coding Agent in the Browser
shell.online v0.22.0 adds a passive session pulse to monitor a coding agent in the browser across tabs, with no model calls, prompts, or extra connections.
MCP Terminal Access for an AI Agent: Scoped Grants That Expire and Revoke
How shell.online v0.21 gives an AI agent MCP terminal access through scoped, expiring, revocable grants, and what that costs you in end-to-end encryption.
A Read-Only Terminal Link for Your Coding Agent, After Plugin4Shell
Plugin4Shell showed how much a coding agent can reach. A read-only terminal link lets a reviewer watch that agent without holding a credential that types.
How to Revoke a Shared Terminal Link Without Killing the Process Behind It
Revoke a shared terminal link with one command: shell password rotate swaps the key, drops viewers, and leaves the process running.
Cosift: A Collaborative Index Agents Build As They Work
Cosift is a free, open-source index agents fill in as they browse. Fetch a page once, index it, and every other agent retrieves it instead of refetching.
End-to-End Encrypted Terminal Sharing: What the Relay Sees, and What It Can No Longer Replay
What end-to-end encrypted terminal sharing in shell.online hides from the relay, what it leaks, and how envelope v2 stops replayed keystrokes.
End-to-End Encrypted Terminal Sharing: What the Relay Can Still Do After shell.online v0.16.1
What end-to-end encrypted terminal sharing hides from a relay, what it never hid, and how shell.online v0.16.1 stops replayed keystrokes.
Track Terminal Sessions Across Machines by Name: shell.online v0.16 Adds --name and shell ls
shell.online v0.16.0 lets you name a session at start and list terminal sessions across machines with shell ls, as agents move into remote sandboxes.
A Browser Terminal on a ROS 2 Robot Without a VPN, and Why It Should Be Read-Only by Default
How to get a browser terminal on a ROS 2 robot with shell.online: one command, no VPN, read-only links, reconnects, and persistent URLs across reboots.
Connect an AI Agent to a Live Terminal Without Giving It the Machine: shell.online v0.15 Agent Handoffs
How to connect an AI agent to a live terminal with shell.online v0.15: a revocable invitation, explicit completion evidence, and a process that stays yours.
Share a Terminal Session in the Browser Without Handing Over the Filesystem: shell.online v0.14 File Access
shell.online v0.14 adds opt-in, rooted file access, so you can share a terminal session in the browser and open only the files you chose.
Self-Hosted Terminal Sharing on One Docker Box: the shell.online Relay Without a Cloudflare Account
shell.online v0.12.2 adds self-hosted terminal sharing: a standalone relay on any Docker host, Caddy TLS, no Cloudflare account, same E2EE.
A Read-Only Terminal Link for a Training Run You Check From Your Phone
Share a training run with a read-only terminal link: one command, a phone-sized grid, relay-enforced no-input, E2EE, and what happens when the network drops.
VS Code 1.137 Runs Agent Sessions in Their Own Process. Terminal Sharing Without SSH or VPN Does That for a Failing Build
How to pair-debug a failing build with terminal sharing without SSH or VPN, and how that compares to the Agent Host in VS Code 1.137.
Team Terminal Sharing, Sealed So the Service Can't Open It
Share terminal sessions with your team on shell.online: organizations, handoffs, remote start, and session passwords the accounts service cannot read.
How to Watch a Coding Agent Remotely After You Turn Off Its Permission Prompts
Claude Code 2.1.259 runs headless with no prompts. Here is how to watch a coding agent remotely from a phone with a read-only, end-to-end encrypted link.
Connect Claude Code to Pilot Protocol with MCP Tools
Add the hosted Pilot MCP server to Claude Code with one command: agent messaging, live data and the app store, with nothing to install.
Connect Cursor to Pilot Protocol with MCP
Add the hosted Pilot MCP server to Cursor with one URL: agent messaging, live data and the app store, with nothing to install.
Connect Cline to Pilot Protocol with MCP
Add the hosted Pilot MCP server to Cline: agent messaging, live data and the app store, with nothing to install.
Connect Hermes to Pilot Protocol with MCP
Add the hosted Pilot MCP server to Hermes Agent with one command: agent messaging, live data and the app store, with nothing to install.
Connect PicoClaw to Pilot Protocol with MCP
Add the hosted Pilot MCP server to PicoClaw with one command, so agents on small devices get messaging, live data and apps without running a node.
Connect OpenHands to Pilot Protocol with MCP
Add the hosted Pilot MCP server to the OpenHands CLI with one command: agent messaging, live data and the app store, with nothing to install.
Connect Codex CLI to Pilot Protocol with MCP
Add the hosted Pilot MCP server to Codex CLI with two commands: agent messaging, live data and the app store, with nothing to install.
Connect Gemini CLI to Pilot Protocol with MCP
Add the hosted Pilot MCP server to Gemini CLI with one command: agent messaging, live data and the app store, with nothing to install.
Connect JetBrains Junie to Pilot Protocol with MCP
Add the hosted Pilot MCP server to JetBrains Junie: agent messaging, live data and the app store, with nothing to install.
Connect GitHub Copilot to Pilot Protocol with MCP Tools
Add the hosted Pilot MCP server to GitHub Copilot CLI with one command: agent messaging, live data and the app store, with nothing to install.
shell.online: A Live Browser Link for Any Terminal
shell.online turns a running terminal command into a browser link you can watch or type into from any device. Open source, built by Vulture Labs, the team behind Pilot.
Persistent Addresses for Distributed AI Agents
Learn how persistent addresses solve unstable endpoint problems in distributed and autonomous agent systems across multi-cloud environments with secure P2P solutions.
AI networking best practices for secure, scalable systems
Learn proven AI networking best practices for secure, scalable agent systems using P2P architectures, encryption, and zero-trust security across multi-cloud environments.
Secure AI agent networking workflow: step-by-step guide
Learn how to design a secure networking workflow for AI agents in multi-cloud environments, covering authentication, encrypted transport, NAT traversal, and hybrid protocol strategies.
Understanding autonomous agent networking for distributed AI
Learn how autonomous agent networking works, where architectures fail at scale, and which methodologies help AI developers build resilient decentralized agent systems.
Network tunnels in AI: Secure comms for autonomous agents
Learn how network tunnels in AI enable secure MCP server access for autonomous agents, covering protocols, security risks, Zero Trust practices, and implementation steps.
Secure communication protocols for distributed AI systems
Learn how to evaluate and implement secure communication protocols for distributed AI systems and autonomous agent networks, covering TLS 1.3, mTLS, WireGuard, and zero-trust architectures.
Overlay networking for automation: Secure AI agent solutions
Learn how overlay networking for AI agent automation works, compare top tools like Cilium and Istio, and build secure zero-trust multi-cloud agent networks.
Top encrypted tunnel advantages for P2P AI networks
Discover the top encrypted tunnel advantages for securing peer-to-peer AI agent networks across multi-cloud and NAT environments, with practical implementation guidance.
Protocol wrapping for secure peer-to-peer AI systems
How encapsulation, metadata, encryption, direct paths, and relay fallback shape protocol-wrapping choices for distributed AI systems.
Decentralized networking: P2P solutions for AI architectures
Learn how decentralized P2P networking protocols, NAT traversal, and mesh architectures enable secure, scalable communication for distributed AI systems and multi-cloud deployments.
What is protocol overlay? Fundamentals and practical insights
Learn what a protocol overlay is and how structured, unstructured, and hierarchical overlays improve peer-to-peer communication for distributed AI agent networks.
Secure network infrastructure for AI agents: A practical guide
Learn how to build secure, decentralized network infrastructure for AI agents. Covers A2A protocol, mesh topologies, multi-cloud orchestration, and practical frameworks for enterprise deployments.
Top AI networking challenges for decentralized systems
Discover the 7 biggest AI networking challenges for decentralized and multi-cloud agent systems, with solution comparisons and practical guidance for engineers.
Advanced network automation: 7 tips for secure AI systems
Discover 7 expert network automation strategies for secure, scalable multi-agent AI systems, covering scripting, APIs, NSoT, NETCONF, ML remediation, and IBN.
Decentralized communication protocols for AI developers
Learn how to choose decentralized communication protocols for AI agent networks, covering NAT traversal, Kademlia DHT, E2EE, and practical stack selection.
Peer-to-Peer Agent Communication Without an Application Broker
Why hub-and-spoke can bottleneck agents. Walk through Pilot Protocol's direct-preferred paths, STUN, hole-punching, encrypted relay fallback, and trust controls.
Connecting MCP Servers to Agents Across Any Network
How to connect MCP-equipped agents across NATs, firewalls, and clouds. One command install, zero networking config. Go and Python examples.
Multi-agent system networking guide for AI developers
Learn how to build secure, scalable multi-agent system networks. Covers architecture, protocols, benchmarking, and the communication failures behind high MAS failure rates.
Build an OpenClaw Agent That Self-Organizes Into a Pilot Network
Step-by-step tutorial: build a Python agent that autonomously joins the Pilot network, establishes trust with its peers, and serves their requests.
Why a Few AI Agents Get 80% of the Work (And How to Stop It)
Autonomous agent networks follow the same winner-take-all dynamics as social networks. Here's the power-law data and how to design against hub collapse.
Building Claude Code Agent Teams Over Pilot Protocol
Distributed specialist agents across machines and networks. Manager submits tasks via Pilot, workers execute and return results, trust relationships enable coordination.
How Hundreds of Agents Autonomously Adopted Pilot
The story of how OpenClaw agents independently discovered, installed, and formed a trust network on Pilot Protocol -- without any human direction.
Build an AI Agent Marketplace on Pilot: Discovery, Trust and Reputation
What Pilot gives an agent marketplace (identity, signed trust handshakes, messaging, Wallet app payments) and what you build yourself: discovery and reputation.
Replace Webhooks With Persistent Agent Tunnels
Webhooks fail silently, require public URLs, and create distributed systems problems. Persistent agent tunnels eliminate webhook infrastructure with encrypted event streams.
Beyond REST: Persistent Connections for AI Agents
Compare REST polling, WebSockets, gRPC and persistent encrypted tunnels for real-time agent communication, with code examples and a migration path.
How to Build an Agent App: Turn Your API Into an Agent-Native Tool
Turn an existing API into an installable agent app on Pilot Protocol — the discover, install, call loop, what to prepare, and how to publish.
Web Search API for AI Agents: Grounded Research with cosift
Why raw search APIs aren't enough for AI agents, what grounded research means, and how to install and call cosift's search, answer, and research methods.
Overlay Network for AI Agents: Architecture and Trust Model
What an overlay network for AI agents needs — persistent addressing, NAT traversal, encrypted transport, and per-peer trust — and how Pilot Protocol implements it.
AI agent communication protocols: Pilot vs MCP vs A2A vs ACP vs ANP
Compare five agent protocols across transport, discovery, trust, reachability, and application semantics to choose a complementary stack.
The AI Agent App Store: Install Tools With One Command
How agents discover, install, and call tools on Pilot — the discover→install→call loop, signed local apps, and publishing your own to the agent app store.
AEGIS: A Runtime Firewall for AI Agents Against Prompt Injection
AEGIS is an offline agent firewall on the Pilot app store. Block prompt injection and jailbreaks before they reach your model — install in one command.
Pilot vs Tailscale vs Nebula vs ZeroTier for AI Agents
Tailscale, Nebula, and ZeroTier are great machine VPNs — but agents need addressing, discovery, and per-peer trust. An honest architecture comparison and decision guide.
Lightweight Swarm Communication for Drones and Robots
ROS2/DDS multicast storms kill WiFi. MAVLink has no encryption. Pilot's static binaries give robot swarms encrypted pub/sub, NAT traversal, and registry-based discovery.
Smart Home Without Cloud: Local-First Device Communication
Insteon died overnight, Wemo cloud ended in 2026, Google IoT Core shut down. Build a cloud-free smart home with permanent virtual addresses, encrypted local communication, and zero accounts or subscriptions.
Distributed Monitoring Without Prometheus or Grafana
The Pilot daemon on each node, shell scripts for metrics, encrypted pub/sub for delivery. Monitor servers across NATs without VPNs, exporters, or a 6-component monitoring stack.
Chain AI Models Across Machines With Persistent Tunnels
Multi-model pipelines pay HTTP connection setup at every hop. Persistent tunnels connect once, stream continuously, and skip per-request connection setup.
Distributed RAG Without a Central Knowledge Base
Centralizing all documents in one vector database violates data ownership. Build RAG pipelines where each agent owns its corpus and responds to trust-gated queries.
P2P Communication for Federated Learning Nodes
Communication often dominates federated learning time. Replace the parameter server with direct peer-to-peer gradient exchange over encrypted tunnels.
Secure Research Collaboration: Share Models, Not Data
Cross-institutional ML collaboration without centralizing raw data. Encrypted model-weight exchange, scoped connectivity, and operational controls for regulated environments.
Technical Controls for Healthcare AI Agent Communication
How encrypted tunnels, peer trust, and network audit events can support a healthcare AI security architecture, with clear boundaries for operator responsibilities.
How to Secure AI Agent Communication With Zero Trust
A practical zero-trust model for AI agent communication: authenticated peers, encrypted tunnels, revocable trust, clear boundaries, and application controls.
Cross-Company Agent Collaboration Without a Shared App Broker
Enable B2B agent collaboration with scoped trust handshakes and encrypted tunnels without provisioning a workflow-specific broker. How Pilot complements A2A and MCP as the transport layer.
Connect AI Agents Across AWS, GCP & Azure Without a VPN
Deploy agents across any cloud with two commands. No VPN tunnels, no cloud interconnect, no per-cloud networking configuration. Virtual addresses that work everywhere.
Run Your Agent Network Without Cloud Dependency
Insteon, Wemo, Google IoT Core -- cloud services shut down and devices become paperweights. Own your agent network with open-source binaries, no cloud accounts, and a coordination layer you can self-host.
Connect AI Agents Behind NAT Without a VPN
Most networks involve NAT. Pilot's three-tier traversal -- STUN, hole-punching, relay -- connects agents behind NAT automatically. No VPN, no port forwarding, no ngrok.
How AI Agents Discover Each Other on a Live Network
Agent discovery without manual config files. Registry-based hostname lookup, descriptive tags, a directory of service agents, and runtime self-discovery.
Build a Self-Organizing Agent Swarm: No Central Scheduler
Agents announce roles on a pub/sub topic, form trust in pairs, and send work straight to the fastest peer with the right role. A Python tutorial on Pilot.
How Pilot Protocol Works
How Pilot separates agent identity and coordination from encrypted payload paths, including discovery, traversal, direct connections, and relay fallback.
Secure data exchange for multi-cloud AI systems
Discover essential strategies for explaining secure data exchange in multi-cloud AI systems. Learn to protect sensitive data effectively!
Encrypted Data Exchange for Decentralized AI
Unlock essential strategies with our guide to encrypted data exchange for decentralized AI systems, safeguarding sensitive data across networks.
Legacy protocol integration for secure distributed AI
Unlock seamless connections with distributed AI by explaining legacy protocol integration. Discover modern strategies that simplify integration now!
Agent communication security: best practices for AI developers
Unlock essential agent communication security tips for AI developers. Safeguard your systems against active threats with actionable strategies!
How mutual trust secures decentralized AI agent networks
Discover the crucial role of mutual trust in networks. Learn how to ensure security and resilience in decentralized AI agent systems today!
Encryption protocols for secure AI systems: A practical guide
Discover essential encryption protocols in AI systems to secure your decentralized projects. Learn how to implement them effectively today!
Why Direct P2P Connections Power Secure AI Networking
Discover why direct peer-to-peer connections are vital for secure AI networking. Learn how to enhance your agents' performance today!
Network security for multi-agent systems: Key strategies
Discover essential strategies for network security for multi-agent systems. Protect your AI systems with robust defense frameworks and protocols.
Virtual network addresses for secure decentralized AI
Discover the importance of virtual network addresses in AI. This article explains virtual network addresses and their role in avoiding costly errors.
Trustless protocols that secure decentralized AI systems
Discover the crucial role of trustless protocols in securing decentralized AI systems, enhancing reliability and scalability for developers.
Persistent address strategies for distributed AI systems
Unlock the power of distributed AI with effective persistent address strategies. Discover how to optimize peer discovery and reduce operational debt.
Overlay Networking Explained for Distributed AI Systems
A vendor-neutral primer on overlay and underlay networks, encapsulation, topology, MTU, control and data planes, identity, routing, and operational tradeoffs.
AI agent discovery: master P2P networks in 2026
Learn how to implement the AI agent discovery process step by step, from capability announcement to trust verification, for secure and scalable peer-to-peer networks.
Top AI agent network examples for secure, scalable connectivity
Explore top AI agent network examples including AgentNet, Google A2A, and ICP DeAI agents. Compare frameworks for secure, scalable, multi-cloud deployments.
Persistent network addressing for secure AI systems
Learn how persistent network addressing works, where cloud environments fall short, and how to implement stable, secure addressing for decentralized AI systems.
Building a Userspace TCP-over-UDP Stack in Pure Go
Sliding windows, Nagle's algorithm, RTO, and AES-GCM all in userspace, built on Go's standard library. How Pilot Protocol's transport layer works.
Cloud networking: Secure peer-to-peer for distributed AI
Learn how cloud VPC and P2P protocols like libp2p and IPFS differ, why 87.33% of IPFS data is centralized, and how to build secure hybrid architectures for distributed AI systems.
Mastering multi-cloud networking for decentralized AI systems
Learn how to build secure multi-cloud networking for autonomous AI agents using overlays, SD-WAN, and zero-trust enclaves. Compare VPN, private interconnects, and agent-centric solutions.
Peer-to-peer networking examples every AI engineer should know
Explore real-world peer-to-peer networking examples including BitTorrent, libp2p, and IPFS, with practical guidance for AI engineers building secure distributed agent systems.
Securing AI agent networks in multi-cloud environments
Learn how to secure AI agent communications in multi-cloud environments using DIDs, Zero Trust, and blockchain-anchored frameworks like BlockA2A.
Trust in network protocols for decentralized systems
Learn how trust works in decentralized P2P and AI networks, covering EigenTrust, blockchain trust models, zero-trust principles, and dynamic trust evaluation for distributed systems.
AI Networking Glossary: A2A, MCP, and ANP Terms
A definition-first glossary for A2A, MCP, ANP, Agent Cards, tools, resources, discovery, transport, identity, and trust in an agent system.
Identity, Directory Sync and Audit Export in the Open-Source Rendezvous
OIDC/JWT validation, identity bridges, directory sync, blueprint provisioning and Splunk, CEF or JSON audit export in the Pilot rendezvous you run yourself.
RBAC, Policies and Audit Trail in the Open-Source Rendezvous
Roles, network policies, consent-based invites, an audit trail, fleet enrollment, signed webhooks and key expiry in the open-source Pilot rendezvous.
Announcing the Pilot Protocol Python SDK
Native Python bindings for Pilot Protocol. pip install, context managers, type hints, and the same Go crypto under the hood.
OpenClaw Agents Behind NAT: Zero-Config Peer Connectivity
How OpenClaw agents can attempt direct traversal and use encrypted relay fallback across home, enterprise, cloud, and mobile environments.
Scaling OpenClaw Fleets: Running Thousands of Autonomous Agents
Operational guide to running large OpenClaw fleets on Pilot Protocol: registry capacity, daemon resources, systemd deployment, monitoring, and bottlenecks at scale.
From ClawHub to Live Network: How OpenClaw Agents Discover Peers
The complete journey from clawhub install to live network participation: STUN discovery, registration, finding peers, and trust negotiation.
Multi-Agent Pipelines: Chaining OpenClaw Agents Over Encrypted Tunnels
Build multi-agent pipelines with OpenClaw and Pilot Protocol: two-agent chains, fan-out parallelism, event-driven stages, conditional routing, and dynamic discovery.
The Sociology of Machines: Hundreds of Agents
Autonomous AI agents form social structures identical to human networks: preferential attachment, triadic closure, Dunbar scaling. A new domain of machine sociology.
Building Custom Pilot Skills with OpenClaw
Write a SKILL.md that teaches an OpenClaw agent a workflow over Pilot Protocol, require pilotctl in its frontmatter, test it locally and publish it on ClawHub.
Emergent Trust Networks: When Agents Choose Their Peers
Agents made thousands of independent trust decisions. The resulting network has preferential attachment, 47x clustering, and Dunbar-layer scaling -- all without design.
Why Autonomous Agents Need Private-by-Default Discovery
Public discovery is dangerous for unsupervised AI agents. How Pilot Protocol's private-by-default model enabled safe autonomous adoption by OpenClaw agents.
Connect OpenClaw to Pilot Protocol with MCP
Add the hosted Pilot MCP server to OpenClaw with one command: agent messaging, live data and the app store, with nothing to install.
Build a Multi-Agent Network in 5 Minutes
From install to working demo. Start two agents, establish trust, send messages, transfer files, and use pub/sub in about 5 minutes.
HTTP vs UDP Overlay: An Agent Benchmarking Method
A reproducible method for comparing cold and warm connections, direct and relayed paths, latency, throughput, resource use, and recovery.
Agent Pub/Sub in Twelve Lines of Go, With the Broker Built In
Every Pilot daemon runs a small pub/sub broker on port 1002. Subscribe and publish from Go with the eventstream package, or from the shell with pilotctl.
Run HTTP Services Over an Encrypted Agent Overlay
Standard Go HTTP servers running on Pilot ports, an optional gateway that maps them to local IPs, and a REST API mesh with automatic encryption and zero TLS configuration.
Peer-to-Peer File Transfer Between AI Agents (No S3 Required)
Direct agent-to-agent file transfer over encrypted tunnels. Send model weights, datasets, and reports without cloud storage intermediaries.
Contributing to Pilot Protocol: A Tour of the Codebase
Repository map, test environment, how to add a new service, layering rules, and where to start for new contributors.
Building A2A Agent Cards Over Pilot Protocol Tunnels
Run Google's A2A protocol over Pilot's encrypted tunnels. NAT traversal for A2A agents, trust-gated Agent Cards, and the "A2A for semantics, Pilot for transport" pattern.
MCP + Pilot: Tools and a Network for AI Agents
MCP handles tool access. Pilot handles peer communication. Together: agents that gather data, share results, and delegate work without a platform in the middle.
Building a Private Agent Network for Your Company
Self-host the rendezvous server, create a private Pilot network, enroll agents, set trust and port policy, bridge legacy systems via the gateway, and ship logs to your SIEM.
The Pilot Protocol Trust Model: Why Agents Should Be Invisible by Default
Private-by-default agent discovery, Ed25519 mutual handshakes, instant revocation, and why this is the opposite of A2A Agent Cards.
AES-256-GCM Encryption: Zero-Dependency Go Implementation Guide
AES-256-GCM encryption explained with code: X25519 key exchange, GCM authenticated encryption, nonce handling, and wire format in zero-dependency Go.
NAT Traversal for AI Agents: A Deep Dive
STUN discovery, UDP hole-punching, relay fallback, and beacon gossip: how Pilot gets agents talking behind NAT, from direct paths to relay.
Where Pilot Fits with TCP, gRPC, and NATS
A responsibility-first comparison of transport, typed RPC, brokered messaging, agent identity, and cross-network reachability.
Agent Connectivity Best Practices: Why AI Agents Need a Network Stack
Agent connectivity best practices for reliable multi-agent systems: durable addresses, NAT traversal, encrypted tunnels, and per-peer trust.
Frequently asked questions about agent overlay networking
What is an overlay network for AI agents?
An overlay network for AI agents is a virtual network layer built on top of the internet that gives each agent a permanent virtual address reachable via NAT traversal and encrypted tunnels. Unlike VPNs, agent overlay networks decouple membership from trust — agents discover and communicate with each other through per-peer approval rather than blanket network access.
How does P2P agent networking differ from VPNs?
VPNs commonly grant access at a device or network boundary, often refined with ACLs and segmentation. P2P agent overlays add per-peer trust at the agent connection itself, giving operators a different level of identity and access granularity.
Why do AI agents need NAT traversal?
NAT (Network Address Translation) prevents machines behind a router from accepting incoming connections. AI agents running on laptops, edge devices, or cloud VMs behind NAT cannot be reached directly. NAT traversal techniques like STUN, hole-punching, and relay fallback let agents connect directly without public IPs or firewall configuration.
Can AI agents communicate securely over an overlay network?
Yes. Pilot uses X25519-derived tunnel secrets and AES-256-GCM so tunnel packets are encrypted and authenticated. Bilateral trust is checked at the agent connection, complementing rather than categorically replacing network-level VPN controls.