worldarchitect-codebase-sherpa

name	worldarchitect-codebase-sherpa
description	Orientation guide for navigating the WorldArchitect.ai repository and tooling
type	usage
scope	project

WorldArchitect Codebase Sherpa

Purpose

Equip Claude with a quick-start orientation for the WorldArchitect.ai repository so it can answer navigation, onboarding, and architecture questions without deep prompting.

When to activate

• A user asks for codebase orientation, file locations, or architectural explanations.
• Someone needs environment spin-up steps or CLI references.
• A request mentions "where is", "how do I run", "what does" regarding repository assets.

Repository tour

• Core application: mvp_site/ houses the Flask backend, MCP servers, templates, static assets, and tests. Frontend variants live in mvp_site/frontend_v1/ and mvp_site/frontend_v2/.
• Game content: Narrative assets in world/ and world_reference/. Campaign prompts live under prompt_archive/.
• Automation: Scripts for local dev and CI under scripts/ plus top-level run_*.sh helpers.
• Documentation: docs/, roadmap/, analysis/, and GENESIS.md capture requirements, planning notes, and retros.
• Agent tooling: .claude/, claude_command_scripts/, and mcp_servers/ define MCP behaviors and bot orchestration.

Quick spin-up checklist

1. Install deps: pip install -r requirements/base.txt
2. Environment: Copy .env.example to .env; ensure TEST_MODE=mock for local work.
3. Run API locally:

   ./vpython mvp_site/main.py serve
   # or
   ./run_local_server.sh
   

4. Pair MCP tools: Use ./scripts/install_mcp_servers.sh claude or ./start_game_mcp.sh depending on the workflow.
5. Stop services: ./kill_dev_server.sh cleans up hanging processes.

Frequently referenced commands

• Tests: ./run_tests.sh, ./run_tests.sh --integration, ./run_tests_with_coverage.sh
• Linting: ./run_lint.sh, pre-commit run -a
• UI checks: ./run_ui_tests.sh
• Deployment helpers: ./deploy.sh, ./deploy_mcp.sh
• Oracle CLI: source scripts/oracle_helpers.sh then use helpers like oracle_arch_preview (dry-run bundle), oracle_arch (architecture review), oracle_ai_debug (AI pipeline bug with optional report path), oracle_diff_review (reviews git diff), oracle_ui_debug (frontend triage). Playbook at docs/oracle_playbook.md.

Architectural highlights

• MCP-centric: Backend routes in mvp_site/main.py broker requests to modular MCP servers under mcp_servers/.
• Persona system: Narrative logic leans on persona configs in world/ and selectors defined in mvp_site/services/persona_service.py.
• Persistence: Firestore integrations live in firestore_service.py and mvp_site/services/firestore/ modules.
• Testing layout: Unit tests reside beside code in mvp_site/tests/; cross-system flows sit in mvp_site/test_integration/ and top-level tests/.

Knowledge links

• Company vision and values: docs/vision/ and docs/company/.
• Gameplay overview: docs/product/game_loop.md and GENESIS.md.
• Roadmap context: roadmap/ for upcoming milestones.

Response style

• Provide directory references with relative paths.
• Suggest the minimal command sequence needed.
• Surface related docs when they clarify the task.