Open-source agent network lets agents find paid work
A new open-source project lets AI agents register on a network, discover tasks via the Agent Working Protocol, and autonomously earn money for completed work. The idea signals growing infrastructure to treat agents as first-class, monetizable actors in distributed systems. (x.com)
AWP’s public testnet console lists 223 agents registered, one active subnet, and “current_epoch 0” in its live status output. (awp.pro/testnet) The AWP RootNet implementation runs on BNB Smart Chain (BSC) and documents that subnets auto-deploy a SubnetManager with Merkle-based reward distribution and configurable AWP strategies such as Reserve, AddLiquidity, and BuybackBurn. (github.com/awp-core/awp-skill-test/blob/main/SKILL.md, awp.pro) The project’s Benchmark Subnet is explicitly billed as the “first subnet” on AWP and crowdsources AI benchmark datasets by having agents both submit discriminative questions and answer peers’ questions, with rewards computed by question quality and answer accuracy. (github.com/awp-core/subnet-benchmark) The Benchmark Subnet repository describes a daily settlement that calculates final rewards and publishes a Merkle tree for on‑chain token claiming, and the README states a planned mainnet deployment on BSC plus Base. (github.com/awp-core/subnet-benchmark) An OpenClaw skill maintained for AWP exposes agent-facing operations — the skill lists 22 actions covering protocol queries, registration, staking, subnet management, governance proposals and voting, plus real‑time WebSocket monitoring of 27 event types. (github.com/awp-core/awp-skill-test/blob/main/README.md) AWP’s public materials and repository READMEs reference a protocol whitepaper and a Proof-of-Useful-Work design as core components of emissions, verification, and governance mechanics under development. (awp.pro, github.com/awp-core/subnet-benchmark) An IETF Internet‑Draft registers a Well‑Known URI /.well‑known/awp.json for an “Agent Workflow Protocol” resource and lists an expiry date of 5 May 2026 on the datatracker entry. (datatracker.ietf.org/doc/draft-vinaysingh-awp-wellknown/00/)
