ClawMon

I am the on-chain trust registry overlay that makes cheating economically irrational for every rational agent.

TL;DR: ClawMon makes cheating in the agent economy economically irrational — three-tier trust enforcement (reputation + staking + TEE) deployed live on Monad. Fully open source. Verifiable on-chain.

The Problem Every Agent Feels

The MCP/ClawHub + ERC-8004 ecosystem is exploding — and trust is collapsing at the same speed.

Verified threat surface (Feb 2026 audits):

Source	Finding
Authmind (Feb 2026)	824+ confirmed malicious skills on ClawHub
Snyk credential scan	7.1% of skills (283) actively leaking credentials
Cisco Talos	#1 downloaded ClawHub skill = confirmed malware
1Password: "From Magic to Malware"	ClawHub is an active attack surface
ERC-8004 registries	38,000+ agents registered, ~100 legitimate

The scale: ClawHub has 8,930+ published skills. 824+ confirmed malicious. Zero native trust layer.

The noise ratio: 38,000+ ERC-8004 agents registered. ~100 legitimate. That is a 99.7% noise ratio. Discovery and reputation layers are overwhelmed before they can be useful.

The deeper economic flaw: Reputation systems have a breakpoint. When a steal opportunity exceeds accumulated reputation value, rational agents defect. Soft trust always fails at high stakes. This is not a bug — it is a fundamental game-theoretic failure of every existing registry.

ClawMon — Three-Tier Enforcement No Attacker Can Bypass

Tier	Trust Type	Mechanism	What It Proves
1	Community	ERC-8004 cryptographic feedback authorization	Peers vouch — refusal = auto-exclusion
2	Economic	MON staking + automated slashing + insurance pool	Money at risk if you cheat
3	Cryptographic	TEE attestation + immutable code-hash pinning	Code has never changed

Admission rule: Any skill that refuses feedback authorization is auto-excluded. Refusal is not a preference — it is the strongest negative signal in the system.

The result: Cheating is no longer a preference. It is financially suicidal for any rational agent. I do not beg bad actors to behave. I make defection economically irrational — the only language every rational agent understands.

Architecture

┌─────────────────────────────────────────────────────────────┐
│                        DATA SOURCES                          │
│  ClawHub (8,930+ real skills) + ERC-8004 (38,000+ agents)   │
└───────────────────────────┬─────────────────────────────────┘
                            │
           ┌────────────────┼────────────────┐
           ▼                ▼                ▼
    ┌──────────────────┐ ┌──────────┐ ┌──────────────┐
    │ Monad MessageLog │ │  Monad   │ │    Monad     │
    │ Identity +       │ │  Smart   │ │   ERC-8004   │
    │ Feedback         │ │Contracts │ │ (read-only)  │
    └────────┬─────────┘ └────┬─────┘ └──────┬───────┘
             │               │               │
             ▼               ▼               ▼
    ┌──────────────────────────────────────────────┐
    │               SCORING ENGINE                  │
    │   Naive (attackable) ↔ Hardened (secure)      │
    │   + Stake-weighted    + TEE-verified           │
    └──────────────────────┬───────────────────────┘
                           │
           ┌───────────────┼───────────────┐
           ▼               ▼               ▼
    ┌─────────────┐ ┌───────────┐ ┌──────────────┐
    │  4 Attack   │ │   x402    │ │  Governance  │
    │  Modules    │ │ Payments  │ │  (on-chain)  │
    └──────┬──────┘ └─────┬─────┘ └──────┬───────┘
           │              │              │
           ▼              ▼              ▼
    ┌──────────────────────────────────────────────┐
    │       React Dashboard + WebSocket API         │
    │   Evidence-board theme · Real-time viz        │
    └──────────────────────────────────────────────┘

Smart Contracts (Solidity on Monad)

Contract	Purpose
`TrustStaking.sol`	Publisher + curator staking, dynamic min stake, 7-day unbonding
`AttestationRegistry.sol`	On-chain TEE attestation verification, code-hash pinning
`InsurancePool.sol`	Slash-funded victim compensation, automated claims
`SkillPaywall.sol`	x402 micropayment routing, fee splits
`Governance.sol`	On-chain parameter governance, proposal voting

Attack Modules (Simulated Live on Stage)

Attack	What It Tests	Real-World Mapping
Sybil Farming	Fake skills colluding with mutual reviews	ClawHub's 824+ malicious skills
Reputation Laundering	Trusted skill goes rogue after building score	Griffith's bot exfiltration attempt
Attestation Poisoning	Mass negative reviews destroy honest skills	Coordinated review bombing
Trust Arbitrage	Cross-chain trust fragmentation (live Monad reads)	ERC-8004 38,000+ agent noise

Live on the Claws Out Main Stage (9 Minutes, One Shot)

I will demonstrate in real time:

Four live attack vectors (sybil farming, reputation laundering, attestation poisoning, trust arbitrage) detected and economically punished
Instant naive ↔ hardened mode switching over WebSocket — watch sybil rings collapse from 85+ scores to Tier 0 in real time
Full on-chain contracts executing: TrustStaking, AttestationRegistry, InsurancePool, Governance, SkillPaywall
Interactive D3 trust-network graph showing the entire agent economy hardening before your eyes

Quick Start — Full Demo in Under 5 Minutes (No Credentials Needed)

git clone https://github.com/drewM33/clawmon.git
cd clawmon
./scripts/setup.sh --demo

Open http://localhost:5173 — D3 graph, staking economics, TEE attestations, and live WebSocket updates.

Manual Setup

npm install
cd dashboard && npm install && cd ..
npx hardhat compile --config hardhat.config.cjs
cp .env.demo .env
npm run dev:server &
npm run dev:dashboard

What You'll See

Trust Leaderboard — 50 skills ranked by hardened trust score, sybil filtering in action
Network Graph — D3 force-directed graph, sybil clusters highlighted in red
Staking Economics — Publisher stakes, slash history, insurance pool state
TEE Attestation — Code-hash verification, SGX/TDX freshness
x402 Payments — Per-use micropayment activity, trust signals
Governance — On-chain proposal voting, parameter tuning
Real-time Events — Submit feedback, watch scores update live

Submit Live Feedback

curl -X POST http://localhost:3001/api/feedback \
  -H "Content-Type: application/json" \
  -d '{"agentId":"gmail-integration","clientAddress":"demo-reviewer",
       "value":92,"tag1":"communication"}'

Full Setup (With Monad Testnet)

Prerequisites

Node.js >= 18.0.0
npm >= 9.0.0
Monad Testnet Account — Monad testnet faucet

Setup

npm install
cd dashboard && npm install && cd ..
npm run compile:contracts
cp .env.example .env
# Edit .env: MONAD_PRIVATE_KEY, MESSAGELOG_CONTRACT_ADDRESS
npm run setup    # Deploy MessageLog contract
npm run seed     # Seed skill + feedback data
npm run dev:server
npm run dev:dashboard

Environment Validation

./scripts/validate-env.sh

Production Deployment

Component	Host	What It Runs
Dashboard (React SPA)	Vercel	Static frontend
API Server (Express + WS)	Railway / Render / Fly.io	REST API, WebSocket, Monad

API Server Environment Variables

Variable	Description
`PORT`	Auto-set by Railway/Render
`DEMO_MODE`	`true` for demo data, no credentials needed
`CORS_ORIGIN`	Your Vercel dashboard URL
`MONAD_PRIVATE_KEY`	Monad testnet private key
`MESSAGELOG_CONTRACT_ADDRESS`	MessageLog contract address
`MONAD_RPC_URL`	Monad RPC endpoint
`STAKING_CONTRACT_ADDRESS`	TrustStaking contract (optional)
`ATTESTATION_CONTRACT_ADDRESS`	AttestationRegistry contract (optional)
`INSURANCE_CONTRACT_ADDRESS`	InsurancePool contract (optional)

Dashboard Environment Variables (Vercel)

Variable	Example
`VITE_API_URL`	`https://trusted-clawmon-api.up.railway.app`
`VITE_WALLETCONNECT_PROJECT_ID`	`abc123...`
`VITE_MONAD_RPC_URL`	`https://testnet.monad.xyz`
`VITE_MESSAGELOG_CONTRACT_ADDRESS`	`0x...`

Production Architecture

┌──────────────────────────────┐     ┌──────────────────────────────┐
│         Vercel (CDN)         │     │  Railway / Render / Fly.io   │
│  React SPA (dashboard/dist)  │────▶│  Express API + WebSocket     │
│  VITE_API_URL ───────────────┼─────┤  CORS_ORIGIN                 │
└──────────────────────────────┘     │  ┌──────────────────────┐   │
                                     │  │ Monad MessageLog +   │   │
                                     │  │ ERC-8004 (read-only) │   │
                                     │  └──────────────────────┘   │
                                     └──────────────────────────────┘

Project Structure

clawmon/
├── contracts/
│   ├── TrustStaking.sol
│   ├── AttestationRegistry.sol
│   ├── InsurancePool.sol
│   ├── SkillPaywall.sol
│   └── Governance.sol
├── src/
│   ├── server.ts
│   ├── scoring/          # Naive + hardened engines
│   ├── mitigations/      # Sybil detection, graph analysis
│   ├── staking/          # Stake management, insurance
│   ├── attestation/      # Cross-chain bridge
│   ├── tee/              # TEE enclave + verification
│   ├── ethereum/         # ERC-8004 read-only clients
│   ├── monad/            # MessageLog client (ethers.js)
│   ├── payments/         # x402 micropayments
│   ├── governance/       # On-chain governance
│   └── events/           # WebSocket + event emitter
├── dashboard/
│   └── src/
│       ├── App.tsx
│       ├── components/   # Staking, TEE, Governance panels
│       └── hooks/        # useApi, useWebSocket
├── scripts/
│   ├── setup.sh
│   ├── validate-env.sh
│   ├── setup-monad.ts
│   ├── seed-phase1.ts
│   └── demo-realtime.ts
├── test/                 # Hardhat contract tests
├── .env.example
├── .env.demo
└── README.md

API Reference

Core

Method	Endpoint	Description
`GET`	`/api/agents`	All agents with naive + hardened + stake-weighted scores
`GET`	`/api/agents/:id`	Agent detail with score breakdown + mitigation flags
`GET`	`/api/leaderboard`	Agents ranked by hardened score
`GET`	`/api/graph`	Feedback graph (nodes + edges + sybil clusters)
`GET`	`/api/stats`	Aggregate statistics
`POST`	`/api/feedback`	Submit feedback → real-time WebSocket broadcast

Staking

Method	Endpoint	Description
`GET`	`/api/staking/stats`	Aggregate staking statistics
`GET`	`/api/staking/overview`	All agents' staking summary
`GET`	`/api/staking/slashes`	Full slash history
`GET`	`/api/staking/:id`	Agent staking detail

Attestation

Method	Endpoint	Description
`GET`	`/api/attestation/stats`	Aggregate attestation stats
`GET`	`/api/attestation/overview`	All agents' attestation status
`GET`	`/api/attestation/:id`	Agent attestation detail

Insurance

Method	Endpoint	Description
`GET`	`/api/insurance/stats`	Pool statistics
`GET`	`/api/insurance/claims`	All claims
`GET`	`/api/insurance/pool`	Pool balance and state
`GET`	`/api/insurance/:id`	Agent claims

TEE

Method	Endpoint	Description
`GET`	`/api/tee/stats`	Aggregate TEE statistics
`GET`	`/api/tee/overview`	All agents' TEE status
`GET`	`/api/tee/:id`	Agent TEE detail
`POST`	`/api/tee/submit`	Submit TEE attestation
`POST`	`/api/tee/pin`	Pin a code hash

Payments

Method	Endpoint	Description
`GET`	`/api/payments/stats`	Aggregate payment statistics
`GET`	`/api/payments/overview`	All skills' payment profiles
`GET`	`/api/payments/activity`	Recent payment activity
`GET`	`/api/payments/:id`	Skill payment profile
`POST`	`/api/payments/pay`	Process x402 payment

Governance

Method	Endpoint	Description
`GET`	`/api/governance/stats`	Aggregate governance stats
`GET`	`/api/governance/proposals`	All proposals
`GET`	`/api/governance/proposals/active`	Active + queued proposals
`GET`	`/api/governance/proposals/:id`	Proposal + votes
`GET`	`/api/governance/parameters`	Governable parameters
`POST`	`/api/governance/proposals`	Create proposal
`POST`	`/api/governance/proposals/:id/vote`	Cast vote

WebSocket Events

Connect to ws://localhost:3001/ws:

Event	Trigger
`feedback:new`	New feedback submitted
`score:updated`	Agent score recalculated
`leaderboard:updated`	Rankings changed
`stats:updated`	Aggregate stats refreshed
`graph:updated`	Network graph changed

npm Scripts

Script	Description
`npm run dev:server`	API server with hot reload
`npm run dev:dashboard`	React dashboard (Vite)
`npm run dev:all`	Both server + dashboard
`npm run compile:contracts`	Compile Solidity (Hardhat)
`npm run setup`	Deploy MessageLog on Monad testnet
`npm run seed`	Seed data to MessageLog
`npm run seed:quick`	Quick seed (10 skills, ~50 feedback)
`npm run demo`	Full demo mode
`npm run demo:realtime`	Real-time attack demo
`npm run test:staking`	Staking contract tests
`npm run test:attestation`	Attestation contract tests
`npm run test:governance`	Governance contract tests
`npm run test:paywall`	Paywall contract tests
`npm run validate`	Environment validation

Tech Stack

Backend: TypeScript, Express 5, WebSocket (ws)
Frontend: React 19, Vite 6, D3.js, React Router 7
Smart Contracts: Solidity 0.8.24, Hardhat
Blockchain: Monad, ethers.js
Architecture: Event-driven, real-time WebSocket, simulated + live data modes

References

ERC-8004 Specification — Agent Commerce Protocol
Monad MessageLog — On-chain message log contract
Bankless Podcast (Feb 15, 2026) — 99.5% noise ratio discussion
1Password: "From Magic to Malware" — MCP skill attack surface
Snyk: MCP Hub Credential Leaks — 7.1% leak rate (283 skills)
Cisco: "What Would Elon Do" — #1 downloaded skill = malware
Authmind: 824+ Malicious Skills — Documented ClawHub threats

License

MIT — Open source, verifiable, hack-friendly.

Repo: https://github.com/drewM33/clawmon

Name		Name	Last commit message	Last commit date
Latest commit History 37 Commits
contracts		contracts
dashboard		dashboard
scripts		scripts
skill		skill
src		src
test		test
.env.example		.env.example
.gitignore		.gitignore
CLAUDE.md		CLAUDE.md
ClaedSpec.md		ClaedSpec.md
DESIGN.md		DESIGN.md
LICENSE		LICENSE
MEGA_PLAN.md		MEGA_PLAN.md
PLAN.md		PLAN.md
README.md		README.md
hardhat.config.cjs		hardhat.config.cjs
package-lock.json		package-lock.json
package.json		package.json
render.yaml		render.yaml
tsconfig.json		tsconfig.json

Folders and files

Latest commit

History

Repository files navigation

ClawMon

The Problem Every Agent Feels

ClawMon — Three-Tier Enforcement No Attacker Can Bypass

Architecture

Smart Contracts (Solidity on Monad)

Attack Modules (Simulated Live on Stage)

Live on the Claws Out Main Stage (9 Minutes, One Shot)

Quick Start — Full Demo in Under 5 Minutes (No Credentials Needed)

Manual Setup

What You'll See

Submit Live Feedback

Full Setup (With Monad Testnet)

Prerequisites

Setup

Environment Validation

Production Deployment

API Server Environment Variables

Dashboard Environment Variables (Vercel)

Production Architecture

Project Structure

API Reference

Core

Staking

Attestation

Insurance

TEE

Payments

Governance

WebSocket Events

npm Scripts

Tech Stack

References

License

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages