# Tiber Fantasy — Multi-Agent Onboarding **Read this file first before starting any task.** --- ## What Is This Project? Tiber Fantasy is a free, open-source NFL fantasy football analytics platform. It provides FORGE-powered player rankings, Data Lab with advanced metrics, Personnel Grouping Intelligence, X Intelligence Scanner, and more. Full architecture details are in `replit.md`. ## Before You Start Any Task 1. **Read `replit.md`** — Full project architecture, schema, conventions, and feature inventory 2. **Read `.claude/conventions.md`** — Coding patterns, naming rules, and guardrails 3. **Check `.claude/context-log.md`** — See what changed recently across all agents 4. **Check your agent log** — Read your own work log in `.claude/agents/` to see your past contributions 5. **Check other agents' logs** — Skim `.claude/agents/` to see what others have done recently 6. **Run `git log --oneline -15`** — Verify the latest state of the codebase ## Agent Work Logs Each agent platform has its own work log tracking contributions and session history: | Agent | Work Log | Role | |-------|----------|------| | **Replit Agent** | `.claude/agents/replit-agent.md` | Primary feature builder, UI/UX, frontend/backend implementation | | **Claude Code** | `.claude/agents/claude-code.md` | Shell-based development, bug fixes, data pipeline work | | **Codex** | `.claude/agents/codex.md` | PR-based contributions via GitHub, audits, refactors | ## After You Complete a Task Every agent must do these steps after finishing work: 1. **Append to `.claude/context-log.md`** — Add a timestamped entry: ``` ### YYYY-MM-DD — [Agent Name]: Brief title - **What changed:** Summary of changes - **Files modified:** List of key files - **Validation:** How it was tested/verified - **Notes:** Anything the next agent should know ``` 2. **Update your agent work log** — Append the task to your `.claude/agents/.md` 3. **Update `replit.md`** if you added new features, changed architecture, or modified conventions 4. **Update `.claude/conventions.md`** if you established new patterns other agents should follow ## Task Specs Detailed task specifications live in `.claude/tasks/`. Each task file contains: - Problem description with root cause analysis - Agent onboarding context (files to read, tables involved) - Solution approach with ranked options - Validation criteria with concrete numbers - Resolution section (added after completion) ## Key Project Files | File | Purpose | |------|---------| | `replit.md` | Project architecture, features, and preferences | | `shared/schema.ts` | All database models (Drizzle ORM) | | `server/routes.ts` | Main Express route registration | | `client/src/App.tsx` | Frontend route registration | | `client/src/components/TiberLayout.tsx` | Sidebar navigation and layout | | `client/src/index.css` | Design system and page styles | ## Git Conventions - Codex works via GitHub PRs on branches named `codex/` - Claude Code and Replit Agent commit directly to `main` - Commit messages should be clear and descriptive - Co-author tags are used when applicable (e.g., `Co-Authored-By: Claude Opus 4.6 `)