This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

• Code in recipe-executor is generated using its own recipes in recipes/recipe_executor/
• To modify code properly, edit the recipe files rather than the generated code directly
• Generate/regenerate code using make recipe-executor-create or make recipe-executor-edit

• Install dependencies: make install (uses uv)
• Lint code: make lint (runs uvx ruff check --no-cache --fix .)
• Format code: make format (runs uvx ruff format --no-cache .)
• Type check: make type-check (runs pyright)
• Run all tests: make test or make pytest
• Run a single test: uv run pytest tests/path/to/test_file.py::TestClass::test_function -v
• Upgrade dependency lock: make lock-upgrade

• Use Python type hints consistently including for self in class methods
• Import statements at top of files, organized by standard lib, third-party, local
• Use descriptive variable/function names (e.g., get_workspace not gw)
• Use Optional from typing for optional parameters
• Initialize variables outside code blocks before use
• All code must work with Python 3.11+
• Use Pydantic for data validation and settings

• Run make to create a virtual environment and install dependencies.
• Activate the virtual environment with source .venv/bin/activate (Linux/Mac) or .venv\Scripts\activate (Windows).

• Run make format to format the code.
• Run make lint to check for linting errors.
• Run make test to run the tests.

This section outlines the core implementation philosophy and guidelines for software development projects. It serves as a central reference for decision-making and development approach throughout the project.

Embodies a Zen-like minimalism that values simplicity and clarity above all. This approach reflects:

• Wabi-sabi philosophy: Embracing simplicity and the essential. Each line serves a clear purpose without unnecessary embellishment.
• Occam's Razor thinking: The solution should be as simple as possible, but no simpler.
• Trust in emergence: Complex systems work best when built from simple, well-defined components that do one thing well.
• Present-moment focus: The code handles what's needed now rather than anticipating every possible future scenario.
• Pragmatic trust: The developer trusts external systems enough to interact with them directly, handling failures as they occur rather than assuming they'll happen.

This development philosophy values clear documentation, readable code, and belief that good architecture emerges from simplicity rather than being imposed through complexity.

• KISS principle taken to heart: Keep everything as simple as possible, but no simpler
• Minimize abstractions: Every layer of abstraction must justify its existence
• Start minimal, grow as needed: Begin with the simplest implementation that meets current needs
• Avoid future-proofing: Don't build for hypothetical future requirements
• Question everything: Regularly challenge complexity in the codebase

• Preserve key architectural patterns: Example: MCP for service communication, SSE for events, separate I/O channels, etc.
• Simplify implementations: Maintain pattern benefits with dramatically simpler code
• Scrappy but structured: Lightweight implementations of solid architectural foundations
• End-to-end thinking: Focus on complete flows rather than perfect components

• Use libraries as intended: Minimal wrappers around external libraries
• Direct integration: Avoid unnecessary adapter layers
• Selective dependency: Add dependencies only when they provide substantial value
• Understand what you import: No black-box dependencies

• Implement only essential endpoints
• Minimal middleware with focused validation
• Clear error responses with useful messages
• Consistent patterns across endpoints

• Simple schema focused on current needs
• Use TEXT/JSON fields to avoid excessive normalization early
• Add indexes only when needed for performance
• Delay complex database features until required

• Streamlined MCP client with minimal error handling
• Utilize FastMCP when possible, falling back to lower-level only when necessary
• Focus on core functionality without elaborate state management
• Simplified connection lifecycle with basic error recovery
• Implement only essential health checks

• Basic SSE connection management
• Simple resource-based subscriptions
• Direct event delivery without complex routing
• Minimal state tracking for connections

• Simple topic-based publisher/subscriber
• Direct event delivery without complex pattern matching
• Clear, minimal event payloads
• Basic error handling for subscribers

• Direct integration with PydanticAI
• Minimal transformation of responses
• Handle common error cases only
• Skip elaborate caching initially

• Simplified queue-based processing
• Direct, focused routing logic
• Basic routing decisions without excessive action types
• Simple integration with other components

• Implement complete end-to-end functionality slices
• Start with core user journeys
• Get data flowing through all layers early
• Add features horizontally only after core flows work

• 80/20 principle: Focus on high-value, low-effort features first
• One working feature > multiple partial features
• Validate with real usage before enhancing
• Be willing to refactor early work as patterns emerge

• Emphasis on integration and end-to-end tests
• Manual testability as a design goal
• Focus on critical path testing initially
• Add unit tests for complex logic and edge cases
• Testing pyramid: 60% unit, 30% integration, 10% end-to-end

• Handle common errors robustly
• Log detailed information for debugging
• Provide clear error messages to users
• Fail fast and visibly during development

When faced with implementation decisions, ask these questions:

1. Necessity: "Do we actually need this right now?"
2. Simplicity: "What's the simplest way to solve this problem?"
3. Directness: "Can we solve this more directly?"
4. Value: "Does the complexity add proportional value?"
5. Maintenance: "How easy will this be to understand and change later?"

Some areas justify additional complexity:

1. Security: Never compromise on security fundamentals
2. Data integrity: Ensure data consistency and reliability
3. Core user experience: Make the primary user flows smooth and reliable
4. Error visibility: Make problems obvious and diagnosable

Push for extreme simplicity in these areas:

1. Internal abstractions: Minimize layers between components
2. Generic "future-proof" code: Resist solving non-existent problems
3. Edge case handling: Handle the common cases well first
4. Framework usage: Use only what you need from frameworks
5. State management: Keep state simple and explicit

# Simple, focused SSE manager that does exactly what's needed
class SseManager:
    def __init__(self):
        self.connections = {}  # Simple dictionary tracking

    async def add_connection(self, resource_id, user_id):
        """Add a new SSE connection"""
        connection_id = str(uuid.uuid4())
        queue = asyncio.Queue()
        self.connections[connection_id] = {
            "resource_id": resource_id,
            "user_id": user_id,
            "queue": queue
        }
        return queue, connection_id

    async def send_event(self, resource_id, event_type, data):
        """Send an event to all connections for a resource"""
        # Direct delivery to relevant connections only
        for conn_id, conn in self.connections.items():
            if conn["resource_id"] == resource_id:
                await conn["queue"].put({
                    "event": event_type,
                    "data": data
                })

# Overly complex with unnecessary abstractions and state tracking
class ConnectionRegistry:
    def __init__(self, metrics_collector, cleanup_interval=60):
        self.connections_by_id = {}
        self.connections_by_resource = defaultdict(list)
        self.connections_by_user = defaultdict(list)
        self.metrics_collector = metrics_collector
        self.cleanup_task = asyncio.create_task(self._cleanup_loop(cleanup_interval))

    # [50+ more lines of complex indexing and state management]

• It's easier to add complexity later than to remove it
• Code you don't write has no bugs
• Favor clarity over cleverness
• The best code is often the simplest

This philosophy section serves as the foundational guide for all implementation decisions in the project.

This section outlines the modular design philosophy that guides the development of our software. It emphasizes the importance of creating a modular architecture that promotes reusability, maintainability, and scalability all optimized for use with LLM-based AI tools for working with "right-sized" tasks that the models can easily accomplish (vs pushing their limits), allow working within single requests that fit entirely with context windows, and allow for the use of LLMs to help with the design and implementation of the modules themselves.

To achieve this, we follow a set of principles and practices that ensure our codebase remains clean, organized, and easy to work with. This modular design philosophy is particularly important as we move towards a future where AI tools will play a significant role in software development. The goal is to create a system that is not only easy for humans to understand and maintain but also one that can be easily interpreted and manipulated by AI agents. Use the following guidelines to support this goal:

(how the agent structures work so modules can later be auto-regenerated)

1. Think “bricks & studs.”

   • A brick = a self-contained directory (or file set) that delivers one clear responsibility.
   • A stud = the public contract (function signatures, CLI, API schema, or data model) other bricks latch onto.

2. Always start with the contract.

   • Create or update a short README or top-level docstring inside the brick that states: purpose, inputs, outputs, side-effects, dependencies.
   • Keep it small enough to hold in one prompt; future code-gen tools will rely on this spec.

3. Build the brick in isolation.

   • Put code, tests, and fixtures inside the brick’s folder.
   • Expose only the contract via __all__ or an interface file; no other brick may import internals.

4. Verify with lightweight tests.

   • Focus on behaviour at the contract level; integration tests live beside the brick.

5. Regenerate, don’t patch.

   • When a change is needed inside a brick, rewrite the whole brick from its spec instead of line-editing scattered files.
   • If the contract itself must change, locate every brick that consumes that contract and regenerate them too.

6. Parallel variants are allowed but optional.

   • To experiment, create sibling folders like auth_v2/; run tests to choose a winner, then retire the loser.

7. Human ↔️ AI handshake.

   • Human (architect/QA): writes or tweaks the spec, reviews behaviour.
   • Agent (builder): generates the brick, runs tests, reports results. Humans rarely need to read the code unless tests fail.

By following this loop—spec → isolated build → behaviour test → regenerate—you produce code that stays modular today and is ready for automated regeneration tomorrow.