3.1 KiB
3.1 KiB
MCP Assistant Rules - [Project Name]
Project Context
[Brief description of what your project does and its main purpose. Keep it concise - 2-3 sentences max.]
Core Vision & Architecture
- Product Goal: [Primary goal of your product]
- Target Platform: [Primary platform(s) - web, mobile, desktop, etc.]
- Architecture: [High-level architecture overview]
- Key Technologies: [Main technologies/frameworks used]
Key Technical Principles
[List 4-6 core technical principles that guide your project]
- Example: Session-based architecture with clear boundaries
- Example: API-first design with versioning from day one
- Example: Security by default - validate all inputs at boundaries
- Example: Observable systems with structured logging
Note: The complete project structure and technology stack are provided in the attached project-structure.md file.
Key Project Standards
Core Principles
[List your fundamental development principles]
- Follow KISS, YAGNI, and DRY - prefer proven solutions over custom implementations
- Never mock, use placeholders, or omit code - always implement fully
- Be brutally honest about whether an idea is good or bad
- [Add project-specific principles]
Code Organization
[Define your code organization standards]
- Keep files under [X] lines - split by extracting utilities, constants, types
- Single responsibility per file with clear purpose
- Prefer composition over inheritance
- [Add language/framework specific organization rules]
[Language] Standards
[Replace with your primary language and its standards]
- Type safety requirements
- Naming conventions (classes, functions, constants)
- Documentation requirements (docstring style, required elements)
- Error handling patterns
Error Handling & Logging
- Use specific exceptions with helpful messages
- Structured logging only - define your logging approach
- [Specify logging categories or patterns]
- Every request needs correlation ID for tracing
API Design
[If applicable - define API standards]
- RESTful with consistent URL patterns
- Version from day one (/v1/, /v2/)
- Consistent response format
- Proper HTTP status codes
Security & State
- Never trust external inputs - validate at boundaries
- [Define session/state management approach]
- [Specify data retention policies]
- Keep secrets in environment variables only
Project-Specific Guidelines
[Add any project-specific guidelines that AI assistants should know]
Domain-Specific Rules
[Add rules specific to your problem domain]
Integration Points
[List key integration points or external services]
Performance Considerations
[Add any performance-critical aspects]
Important Constraints
- You cannot create, modify, or execute code
- You operate in a read-only support capacity
- Your suggestions are for the primary AI (Claude Code) to implement
- Focus on analysis, understanding, and advisory support
Quick Reference
[Add frequently needed information]
- Key commands: [List common commands]
- Important paths: [List critical file paths]
- Documentation links: [Add links to detailed docs]