Skip to content

PLAN.md

Latest commit

 

History

History
194 lines (140 loc) · 6.71 KB

PLAN.md

File metadata and controls

194 lines (140 loc) · 6.71 KB

RPB Deployment Plan -- Jurisdiction + Training

Goal

Deploy the Autonet jurisdiction (DAO on werule + RPB contract) and wire the agent framework so users can flip a switch to start training and earning rewards. Each step is independently observable.

Naming

  • RPB = Recursive Principled Body (the system/pattern)
  • Autonet = the first jurisdiction deploying an RPB
  • The contract stays named Autonet.sol for now (rename is cosmetic, do later)
  • In code comments and docs, refer to it as "the RPB contract"

Phase 1: Deploy Autonet jurisdiction on local hardhat

Codebase: c:\code\dao\trustless-contracts

1.1 Create unified deployment script

Write scripts/deploy-autonet-jurisdiction.js that does the full 3-step TrustlessFactory flow + RPB deployment in one shot:

Step 1 -- Infrastructure: trustlessFactory.deployInfrastructure(timelockDelay=1, arbitrationFee=250) Deploys: Economy, Timelock, Registry

Step 2 -- DAO Token: trustlessFactory.deployDAOToken(registry, timelock, tokenParams, govParams) Deploys: RepToken ("Autonet Rep", "AREP"), HomebaseDAO ("Autonet DAO") Initial members: [deployer] with 1000 AREP

Step 3 -- Configure & Finalize: trustlessFactory.configureAndFinalize(addresses, economyParams, registryKeys, registryValues) Links all contracts, transfers control to Timelock. Include in registryKeys: "jurisdiction.rpb" (will be populated after RPB deploy)

Step 4 -- Deploy RPB (Autonet.sol): Deploy AutonetFactory with fresh Autonet implementation. Call factory.createAutonet(economy, registry, repToken, timelock, "Autonet Credits", "ATN") Note: deployer must hold RepToken (satisfied by Step 2 initial distribution).

Step 5 -- Configure RPB:

  • autonet.configureEmission(1000e18, 9900, 300) (1000 ATN/epoch, 1% decay, 5min epochs for testing)
  • Register training service: autonet.registerService(serviceId, economyAddress, gitHash)
  • Activate service: autonet.activateService(serviceId)

Step 6 -- Store RPB address in Registry: Since Registry is now owned by Timelock, either: (a) Include RPB address registration in Step 3's registryKeys (before finalize), or (b) Use Timelock to execute registry.editRegistry("jurisdiction.rpb", rpbAddress) post-deploy. Option (a) requires deploying RPB before finalize. Reorder: Steps 1,2,4,3,5. Option (b) requires a governance proposal. For local testing, deployer IS timelock, so direct call works.

Output: Save all addresses to deployments/autonet-jurisdiction-{network}.json and copy relevant addresses to c:\code\autonet\deployment-addresses.json.

Observable: DAO deployed, RPB deployed, epoch running, service active. Can verify via hardhat console: autonet.currentEpoch() returns 1.

Phase 2: Wire GovernanceBridge into training pipeline

Codebase: c:\code\autonet

2.1 ABI availability

The Python ContractRegistry in nodes/common/contracts.py loads ABIs from artifacts/contracts/governance/AutonetEconomy.sol/Autonet.json.

Create the directory structure and copy the compiled ABI from c:\code\dao\trustless-contracts\artifacts\contracts\Autonet.sol\Autonet.json.

Add a helper script scripts/copy-abis.sh (or .ps1) that does this copy.

2.2 Wire GovernanceBridge in AutonetService

File: c:\code\autonet\nodes\service.py

Add _init_governance() method that:

  1. Reads blockchain config (rpc_url, private_key, chain_id)
  2. Creates BlockchainInterface
  3. Creates ContractRegistry with deployment addresses
  4. Creates GovernanceBridge
  5. Returns it (or None if config missing)

Change _init_training_feed() to pass governance:

governance = self._init_governance()
self._training_feed = TrainingDataFeed(feed_config, governance=governance)

If governance is None, training still works (offline mode -- already supported).

2.3 Update deployment-addresses.json

After Phase 1 deployment, the addresses file needs:

{
  "AutonetEconomy": "0x...",   // RPB contract address
  "AutonetDAO": "0x...",       // HomebaseDAO address
  ...
}

The Phase 1 script handles this.

2.4 Blockchain config in autonet.yaml

Ensure the config dataclass supports:

blockchain:
  rpc_url: "http://127.0.0.1:8545"
  private_key: "0xac0974..."
  chain_id: 31337

Check existing nodes/common/config.py for the config structure.

Observable: Start AutonetService with blockchain config. Training runs. Hardhat console shows attestUsage transactions. Epoch usage increments.

Phase 3: Agent framework defaults

Codebase: c:\code\autonet\atn\

3.1 Pass private_key through AutonetBridge

File: atn/autonet_service.py line ~169

Add alongside existing rpc_url/chain_id passthrough:

if self.config.private_key:
    self._autonet_config.blockchain.private_key = self.config.private_key

3.2 Hardcode Autonet jurisdiction defaults

In ATN config, set default rpc_url and chain_id for Shadownet (or local). Users who flip the training switch don't need to configure anything. The private_key comes from their wallet connection or env var.

Observable: User starts ATN, flips training switch. Training begins. On-chain attestations flow automatically. No manual config needed.

Phase 4: End-to-end verification

4.1 Integration test

Create tests/test_governance_integration.py:

  • Mock BlockchainInterface (don't need real chain for unit tests)
  • Verify GovernanceBridge created when config has rpc_url + private_key
  • Verify TrainingDataFeed receives governance bridge
  • Verify attest_task_completion() calls registry.attest_usage()

4.2 Manual smoke test

  1. Start hardhat node
  2. Run Phase 1 deploy script
  3. Start AutonetService with blockchain config
  4. Generate synthetic agent data (JSONL files)
  5. Trigger training cycle
  6. Check hardhat console for attestUsage transaction
  7. Fast-forward time: await network.provider.send("evm_increaseTime", [300])
  8. Trigger another attestation (auto-advances epoch)
  9. Call claimParticipantReward() from hardhat console
  10. Verify ATN balance > 0

Implementation order

  1. Phase 1: scripts/deploy-autonet-jurisdiction.js in trustless-contracts
  2. Phase 2.1: Copy ABI artifacts
  3. Phase 2.2: Wire GovernanceBridge in nodes/service.py
  4. Phase 2.4: Verify/update config structure
  5. Phase 3.1: Pass private_key in atn/autonet_service.py
  6. Phase 4.1: Integration test
  7. Phase 4.2: Manual smoke test

Deferred (not in this plan)

  • Renaming Autonet.sol to RPB.sol (cosmetic)
  • EvolutionProposal deployment (not needed until architecture changes)
  • Alignment-weighted rewards (add via CapabilityScorecard later)
  • Multi-jurisdiction support (single jurisdiction first)
  • Earn-only token constraint (governance decision, token contract change)
  • Deploy to Shadownet (do after local hardhat verification)