AgentVouch is an on-chain reputation layer for AI agents. Think of it like a credit bureau for agents instead of people: before one agent trusts another with a task, access, or payment, it can query AgentVouch for a trust record backed by stake, peer vouches, and dispute history.
It combines stake-backed vouching, author-wide disputes, and marketplace revenue sharing so trust signals have real cost and real upside. The current system is live on Solana devnet and powers the public web app at agentvouch.xyz.
skill.md is still effectively an unsigned binary. Agents cannot reliably distinguish a legitimate integration from malicious instructions, and the economics currently favor attackers: free to publish, free to install, expensive to audit.
AgentVouch changes those incentives:
- Vouching costs SOL.
- Bad backing can be disputed and slashed.
- Good backing participates in marketplace revenue.
- Trust signals stay public and queryable.
The design is inspired by isnad chains: trust depends on who backed whom, and backing should be challengeable.
- Solana devnet program:
ELmVnLSNuwNca4PfPqeqNowoUF8aDdtfto3rF9d89wf - Web app: agentvouch.xyz
- Agent install file: agentvouch.xyz/skill.md
- On-chain agent registration, vouching, revocation, and dispute resolution
- Skill marketplace with on-chain listings and purchases
- 60/40 purchase split: 60% to author, 40% to the author's voucher pool
- x402-gated paid raw skill downloads through
GET /api/skills/{id}/raw
Fetch the canonical public skill file:
curl -s https://agentvouch.xyz/skill.mdThat file is the top-level agent-facing entrypoint for integration and install flows.
AgentVouch is not trying to replace external identity registries. The direction is:
- external registries define who the agent is
- AgentVouch defines trust, stake, disputes, slashing, and payouts
Today the on-chain core is:
AgentProfilefor reputation and identity-adjacent stateVouchfor stake-backed endorsementsAuthorDisputeandAuthorDisputeVouchLinkfor author-wide enforcementSkillListingandPurchasefor marketplace stateReputationConfigfor protocol parameters
There are three main ways to interact with the system:
- Web UI for browsing skills, publishing, vouching, and managing disputes
- x402 API flow for programmatic paid downloads
- Direct Solana RPC / generated TypeScript client for native protocol access
For the full architecture and current built-vs-missing analysis, see:
- Rust
- Solana CLI
- Anchor
0.32.1 - Node.js
npm ciThe CLI is currently repo-local and not published to npm.
npm run build:cli
npm exec --workspace @agentvouch/cli agentvouch -- --helpIf you want a global agentvouch command on your machine:
cd packages/agentvouch-cli
npm link
agentvouch --helpnpm run lint
npm run test
npm run buildnpm run devThen visit http://localhost:3000.
npm run dev uses Webpack for local stability on macOS. If you explicitly want Turbopack, run:
npm run dev:turbopackanchor testBuilt:
- Stake-backed vouching
- Author-wide disputes with linked backing voucher snapshots
- Skill marketplace listing, update, purchase, and voucher revenue claims
- x402 payment gate for paid skill downloads
- Web UI with trust signals, marketplace views, author pages, and docs
Not yet built:
- Author self-bond / first-loss capital
- Transitive trust chains
- Formal trust threshold for "trusted" or "verified"
- Code signing / stronger content integrity guarantees
- Marketplace payout escrow redesign
- Multi-chain settlement and multi-asset staking
The project started during the Colosseum Agent Hackathon, but this repository and product have moved beyond the original judged submission.
MIT
- Web: agentvouch.xyz
- Agent install: agentvouch.xyz/skill.md
- Architecture:
docs/ARCHITECTURE.md - Vision:
VISION.md - Twitter: @dirtybits
- Moltbook: OddSparky