Skip to content

bounty-proof.md

Latest commit

 

History

History
236 lines (185 loc) · 8.58 KB

bounty-proof.md

File metadata and controls

236 lines (185 loc) · 8.58 KB

Bounty Proof — AgentSafe + SwarmGuard

Sponsor-by-sponsor evidence for ETHDenver 2026 bounty submissions.

1. Base (Coinbase) — Smart Wallet on Base

What We Built

ERC-4337 smart contract wallet deployed to Base (chain ID 8453). Four Solidity modules handle account abstraction, policy enforcement, agent-driven governance, and provenance recording.

Evidence

Artefact Location
AgentSafeWallet.sol packages/contracts/src/AgentSafeWallet.sol
AgentSafeAccount (ERC-4337) packages/contracts/src/account/
PolicyEngine.sol packages/contracts/src/PolicyEngine.sol
GovernanceExecutor.sol packages/contracts/src/GovernanceExecutor.sol
ProvenanceRegistry packages/contracts/src/provenance/
Deploy script packages/contracts/script/Deploy.s.sol
Foundry tests packages/contracts/test/
Chain config (shared) packages/shared/src/constants/chains.ts

Verification Commands

cd packages/contracts

# Build contracts
forge build

# Run contract tests
forge test -vvv

# Check compilation output
ls out/

UI Proof

  • Dashboard (/dashboard): Shows Base chain ID and contract addresses
  • Integrations (/integrations): Base section with deployed addresses and chain ID badge

2. QuickNode — RPC Provider

What We Built

SwarmGuard agents call Base via QuickNode RPC for real-time block data, transaction simulation, and mempool monitoring. The backend detects whether a QuickNode endpoint is configured and falls back to public RPC.

Evidence

Artefact Location
RPC service apps/backend/src/services/rpc/
Chain constants (RPC URLs) packages/shared/src/constants/chains.ts
Health endpoint GET /health → integrations.quicknode
Backend env example apps/backend/.env.example (QUICKNODE_RPC_URL)

Verification Commands

# Check backend health — look for "quicknode" section
curl http://localhost:4000/health | jq '.integrations.quicknode'

# Expected output (live mode):
# { "mode": "live", "blockNumber": 12345678 }

# Expected output (disabled):
# { "mode": "disabled" }

UI Proof

  • Integrations (/integrations): QuickNode section with Live/Disabled badge, block number display (when live), and raw health JSON

3. Kite AI — AI Summarisation & Risk Analysis

What We Built

Kite AI powers the governance proposal summariser and the scam detection NLP pipeline. When a Kite API key is present, the system calls Kite AI for document summarisation. Without a key, a local stub returns structured risk analysis using keyword extraction.

Evidence

Artefact Location
Kite AI service apps/backend/src/services/agents/
Scam Detector agent apps/backend/src/agents/scamDetector.ts
Governance recommend endpoint POST /api/governance/recommend
Health endpoint GET /health → integrations.kiteAi

Verification Commands

# Check Kite AI status
curl http://localhost:4000/health | jq '.integrations.kiteAi'

# Test governance recommendation (uses Kite for summarisation)
curl -X POST http://localhost:4000/api/governance/recommend \
  -H 'Content-Type: application/json' \
  -d '{"proposalId": "prop-001"}' | jq

# Expected: VoteIntent with recommendation, confidence, reasons, policyChecks

UI Proof

  • Integrations (/integrations): Kite AI section with "Run Kite Summary Test" button
  • Governance (/governance): "Get AI Recommendation" button on each proposal card

4. Nouns / Snapshot — Governance Proposal Parsing

What We Built

GovernanceSafe ingests DAO proposals via Snapshot-like API, parses them for risk signals (treasury impact, governance power shifts, urgency flags), and produces AI-guided vote recommendations with human veto.

Evidence

Artefact Location
Proposal service apps/backend/src/governance/proposals.ts
Mock proposals apps/backend/src/governance/mockProposals.json
Governance routes apps/backend/src/routes/governance.ts
Governance runner apps/backend/src/orchestrator/governanceRunner.ts
GovernanceExecutor.sol packages/contracts/src/GovernanceExecutor.sol
Governance V2 schemas packages/shared/src/schemas/governanceV2.ts

Verification Commands

# List proposals
curl http://localhost:4000/api/governance/proposals | jq '.[0]'

# Get AI recommendation for a proposal
curl -X POST http://localhost:4000/api/governance/recommend \
  -H 'Content-Type: application/json' \
  -d '{"proposalId": "prop-001"}' | jq

# Expected: { recommendation, confidence, reasons, policyChecks }

UI Proof

  • Governance (/governance): Proposal cards with risk analysis
  • Governance (/governance): VoteIntent display with policy checks
  • Governance (/governance): Human veto + auto-vote toggle

5. 0g — Decentralised Storage (Stretch Goal)

What We Built

Architecture prepared for storing provenance receipts (swarm consensus decisions) on 0g decentralised storage. Currently a stub — the ProvenanceRegistry records hashes on-chain, and 0g integration would store the full receipt payload off-chain.

Evidence

Artefact Location
ProvenanceRegistry.sol packages/contracts/src/provenance/
Log storage service apps/backend/src/storage/logStore.ts
Integrations page apps/web/src/app/integrations/page.tsx → 0g section

Status

Stub / planned. On-chain hashes are recorded; full 0g blob storage is the next milestone post-hackathon.

Cross-Cutting Proof

Integration Health Harness

# Start backend first, then run:
pnpm healthcheck

# Validates all 6 API endpoints against Zod schemas
# Exit code 0 = all pass, non-zero = failures

Full Build Verification

pnpm install
pnpm build       # Compiles shared + contracts + web + backend
pnpm lint        # ESLint across all packages
pnpm test        # Runs test scripts (includes build)

Architecture Diagram

┌─────────────┐    ┌─────────────────────────────┐
│  Next.js UI  │◄──►│  Express Backend (port 4000) │
│  (port 3000) │    │                             │
│  /dashboard  │    │  ┌─────────────────────┐    │
│  /defense    │    │  │  SwarmGuard Agents   │    │
│  /governance │    │  │  Sentinel, ScamDet,  │    │
│  /policy     │    │  │  MEV, Liquidation    │    │
│  /integrns   │    │  └────────┬────────────┘    │
└─────────────┘    │           │                  │
                   │  ┌────────▼────────────┐     │
                   │  │  Coordinator Agent   │     │
                   │  │  (Consensus Engine)  │     │
                   │  └────────┬────────────┘     │
                   │           │                  │
                   │  ┌────────▼────────────┐     │
                   │  │  Governance Runner   │     │
                   │  │  (Proposal → Vote)   │     │
                   │  └─────────────────────┘     │
                   └──────────────┬───────────────┘
                                  │
                   ┌──────────────▼───────────────┐
                   │  Base (ERC-4337) Contracts    │
                   │  AgentSafeAccount             │
                   │  PolicyEngine                 │
                   │  GovernanceExecutor            │
                   │  ProvenanceRegistry            │
                   └───────────────────────────────┘

Screenshots Checklist (For Submission)

  • Dashboard with swarm status
  • Defense page: suspicious tx evaluation with agent reports
  • Defense page: consensus BLOCK_TX result
  • Governance page: proposal cards loaded
  • Governance page: AI recommendation with policy checks
  • Governance page: human veto applied
  • Integrations page: all sponsor sections with badges
  • Integrations page: raw health/status JSON expanded
  • Terminal: pnpm healthcheck all green
  • Terminal: forge test -vvv all passing