Contributing Guide

Code changes should preserve the main loop: Project, task, proof, review, reward, share pool, memory, and agent readback.

Local Environment

pnpm install
cp .env.example .env
pnpm dev

Run separately:

pnpm api:dev
pnpm web:dev

Change Rules

Scope	Check together
Project	Project detail, Workbench, AgentTurn, API readback
Validation	launch, task, proof, review, reward, settlement
Order	OrderActionPolicy returned actions
Memory	source, entry, contractHash, agent context
Skill	route, execute, verify, project smoke
Docs	docs/book, README.md, related docs/project-*.md

PR Gates

pnpm check
pnpm build
pnpm api:test:unit
pnpm security:secrets
pnpm security:web
git diff --check

Database or jOOQ changes add:

pnpm api:test:integration

Documentation Requirements

Behavior changes should document:

Content	Requirement
Real entry	Page, API, service, or script
State flow	old state -> new state
Verification command	Command that was actually run
Code evidence	File, class, method, or test
User impact	Who sees the change

Agent Skill Changes

Changes under skills/monopolyfun-agent/ need three layers covered:

route-intentResolve intent and resource

execute-actionCall official API write action

verify-actionRead back persisted state

Project/OpenClaw flows also need the relevant project smoke, with durable findings written under docs/.