diff --git a/AGENTS.md b/AGENTS.md
new file mode 100644
index 0000000..a5a3b8b
--- /dev/null
+++ b/AGENTS.md
@@ -0,0 +1,273 @@
+# AGENTS.md
+
+This file provides guidance to LLM when working with code in this repository.
+
+You're experienced Spark developer.  You're developing custom Python data sources to
+simplify work (read/write) with data in different external systems.  You follow the
+architecture and development guidelines outlined below.
+
+## Project Overview
+
+This repository contains source code of Python data source (part of Apache Spark API) for
+working with data in different external systems in batch and/or streaming manner.
+
+**Architecture**:  There is a separate directory for implementation of a data source for a specific external system. Inside it, simple, flat structure with one data source per file is preferred. Each data source implements:
+- `DataSource` base class with `name()`, `reader`, `streamReader`, `writer()`, and `streamWriter()` methods.
+- Separate writer classes for batch (`DataSourceWriter`, `DataSourceReader`) and streaming (`DataSourceStreamWriter`, `DataSourceStreamReader`).
+- Shared writer or reader logic in a common base class.
+
+### Documentation and examples of custom data sources using the same API
+
+There is a number of publicly available examples that demonstrate how to implement custom Python data sources:
+
+- https://github.com/alexott/cyber-spark-data-connectors
+- https://github.com/databricks/tmm/tree/main/Lakeflow-OpenSkyNetwork
+- https://github.com/allisonwang-db/pyspark-data-sources
+- https://github.com/databricks-industry-solutions/python-data-sources
+- https://github.com/dmatrix/spark-misc/tree/main/src/py/data_source
+- https://github.com/huggingface/pyspark_huggingface
+- https://github.com/dmoore247/PythonDataSources
+- https://github.com/dgomez04/pyspark-hubspot
+- https://github.com/dgomez04/pyspark-faker
+- https://github.com/skyler-myers-db/activemq_pyspark_connector
+- https://github.com/jiteshsoni/ethereum-streaming-pipeline/blob/6e06cdea573780ba09a33a334f7f07539721b85e/ethereum_block_stream_chainstack.py
+- https://www.canadiandataguy.com/p/stop-waiting-for-connectors-stream
+- https://www.databricks.com/blog/simplify-data-ingestion-new-python-data-source-api
+
+More information about Spark Python data sources could be found in the documentation:
+
+- https://docs.databricks.com/aws/en/pyspark/datasources
+- https://spark.apache.org/docs/latest/api/python/tutorial/sql/python_data_source.html
+
+
+## Architecture Patterns
+
+### Data Source Implementation Pattern
+
+Most of data sources follows this structure (see [MsSentinel implementation](https://github.com/alexott/cyber-spark-data-connectors/blob/main/cyber_connectors/MsSentinel.py) as reference):
+
+1. **DataSource class**: Entry point, returns appropriate writers
+   - Implements `name()` class method (returns format name like "cassandra")
+   - Implements `writer()` for batch write operations
+   - Implements `streamWriter()` for streaming write operations
+   - Implements `reader()` for batch read operations
+   - Implements `streamReader()` for streaming read operations
+   - Implements `schema` to return a predefined schema for read operations (it could be automatically generated from the response, but it could be slower compared to predefined schema).
+
+2. **Base Writer class**: Shared write logic for batch and streaming
+   - Extracts and validates options in `__init__`
+   - Implements `write(iterator)` that processes rows and returns `SimpleCommitMessage`
+   - May batch records before sending (configurable via `batch_size` option)
+
+3. **Batch Writer class**: Inherits from base writer + `DataSourceWriter`
+   - No additional methods needed
+
+4. **Stream Writer class**: Inherits from base writer + `DataSourceStreamWriter`
+   - Implements `commit()` (handles successful batch completion)
+   - Implements `abort()` (handles failed batch)
+
+5. **Base Reader class**: Shared read logic for batch and streaming reads
+   - Extracts and validates base options in `__init__`
+   - Implements `partitions()` to distribute reads over multiple executors (if it's possible).  The custom class could be used to specify partition information (it should be inherited from `InputPartition`).
+   - Implements `read` to get data for a specific partition.
+
+6. **Batch Reader class**: Inherits from base reader + `DataSourceReader`.
+   - No additional methods needed
+
+7. **Stream Reader class**:  Inherits from base reader + `DataSourceStreamReader`.
+   - `initialOffset` - returns initial offset provided during the first initialization (or inferred automatically).  The offset class should implement `json` and `from_json` methods.
+   - `latestOffset` - returns the latest available offset.
+
+### Key Design Principles
+
+1. **SIMPLE over CLEVER**: No abstract base classes, factory patterns, or complex inheritance
+2. **EXPLICIT over IMPLICIT**: Direct implementations, no hidden abstractions
+3. **FLAT over NESTED**: Single-level inheritance (DataSource → Writer → Batch/Stream)
+4. **Imports inside methods**: For partition-level execution, import libraries within `write()` methods
+5. **Row-by-row processing**: Iterate rows, batch them, send when buffer full
+
+## Adding a New Data Source
+
+Follow this checklist (use existing sources as templates):
+
+1. Create a new folder with the project skeleton.
+2. Create new file `src/python_datasource_connectors/YourSource.py`
+3. Implement `YourSourceDataSource(DataSource)` with `name()`, `writer()`, `streamWriter()`
+4. Implement base writer class with:
+   - Options validation in `__init__`
+   - `write(iterator)` method with write logic
+5. Implement batch and stream writer classes (minimal boilerplate)
+6. Implement base reader class with:
+   - Options validation in `__init__`
+   - `read(partition)` method with read logic
+   - `partitions(start, end)` method to split data into partitions
+7. Implement batch and stream writer classes (minimal boilerplate)
+8. Add exports to `python_datasource_connectors/__init__.py`
+9. Create test file `tests/test_yoursource.py` with unit tests
+10. Update `README.md` with usage examples and options
+
+### Data Source Registration
+
+Users register data sources like this:
+```python
+from dbx import ZipDCMDataSource
+spark.dataSource.register(ZipDCMDataSource)
+
+# Then use with .format("zipdcm")
+df.read.format("zipdcm").option("...", "...").load()
+```
+
+## 🚨 SENIOR DEVELOPER GUIDELINES 🚨
+
+**CRITICAL: This project follows SIMPLE, MAINTAINABLE patterns. DO NOT over-engineer!**
+
+### Forbidden Patterns (DO NOT ADD THESE)
+
+- ❌ **Abstract base classes** or complex inheritance hierarchies
+- ❌ **Factory patterns** or dependency injection containers
+- ❌ **Decorators for cross-cutting concerns** (logging, caching, performance monitoring)
+- ❌ **Complex configuration classes** with nested structures
+- ❌ **Async/await patterns** unless absolutely necessary
+- ❌ **Connection pooling** or caching layers
+- ❌ **Generic "framework" code** or reusable utilities
+- ❌ **Complex error handling systems** or custom exceptions
+- ❌ **Performance optimization** patterns (premature optimization)
+- ❌ **Enterprise patterns** like singleton, observer, strategy, etc.
+
+### Required Patterns (ALWAYS USE THESE)
+- ✅ **Direct function calls** - no indirection or abstraction layers
+- ✅ **Simple classes** with clear, single responsibilities
+- ✅ **Environment variables** for configuration (no complex config objects)
+- ✅ **Explicit imports** - import exactly what you need
+- ✅ **Basic error handling** with try/catch and simple return dictionaries
+- ✅ **Straightforward control flow** - avoid complex conditional logic
+- ✅ **Standard library first** - only add dependencies when absolutely necessary
+
+### Implementation Rules
+
+1. **One concept per file**: Each module should have a single, clear purpose
+2. **Functions over classes**: Prefer functions unless you need state management
+3. **Direct SDK calls**: Call Databricks SDK directly, no wrapper layers
+4. **Simple data structures**: Use dicts and lists, avoid custom data classes
+5. **Basic testing**: Simple unit tests with basic mocking, no complex test frameworks
+6. **Minimal dependencies**: Only add new dependencies if critically needed
+
+### Code Review Questions
+
+Before adding any code, ask yourself:
+- "Is this the simplest way to solve this problem?"
+- "Would a new developer understand this immediately?"
+- "Am I adding abstraction for a real need or hypothetical flexibility?"
+- "Can I solve this with standard library or existing dependencies?"
+- "Does this follow the existing patterns in the codebase?"
+
+## Development Commands
+
+### Python Execution Rules
+
+**CRITICAL: Always use `poetry run` instead of direct `python`:**
+```bash
+# ✅ CORRECT
+poetry run python script.py
+
+# ❌ WRONG
+python script.py
+```
+
+## Development Workflow
+
+### Package Management
+
+- **Python**: Use `poetry add/remove` for dependencies, never edit `pyproject.toml` manually
+- Always check if dependencies already exist before adding new ones
+- **Principle**: Only add dependencies if absolutely critical
+
+### Setup
+```bash
+# Install dependencies (first time)
+poetry install
+
+# Activate environment
+. $(poetry env info -p)/bin/activate
+```
+
+### Testing
+```bash
+# Run all tests
+poetry run pytest
+
+# Run specific test file
+poetry run pytest tests/test_ds.py
+
+# Run single test
+poetry run pytest tests/test_ds.py::TestXxxxDataSource::test_name
+
+# Run with verbose output
+poetry run pytest -v
+```
+
+### Building
+```bash
+# Build wheel package
+poetry build
+
+# Output will be in dist/ directory
+```
+
+### Code Quality
+```bash
+# Format and lint code (ruff)
+poetry run ruff check cyber_connectors/
+poetry run ruff format cyber_connectors/
+
+# Type checking
+poetry run mypy cyber_connectors/
+```
+
+## Testing Guidelines
+
+- Tests use `pytest` with `pytest-spark` for Spark session fixtures
+- Mock external HTTP calls using `unittest.mock.patch`
+- Test writer initialization, option validation, and data processing logic
+- See `tests/test_ds.py` for comprehensive examples
+
+**Test structure**:
+- Use fixtures for common setup (`basic_options`, `sample_schema`)
+- Test data source name registration
+- Test writer instantiation (batch and streaming)
+- Test option validation (required vs optional parameters)
+- Mock HTTP responses to test write operations
+
+## Important Notes
+
+- **Python version**: 3.10-3.13 (defined in `pyproject.toml`)
+- **Spark version**: 4.0.1+ required (PySpark DataSource API)
+- **Dependencies**: Keep minimal - only add if critically needed
+- **Never use direct `python` commands**: Always use `poetry run python`
+- **Ruff configuration**: Line length 120, enforces docstrings, isort, flake8-bugbear
+- **No premature optimization**: Focus on clarity over performance
+
+## Summary: What Makes This Project "Senior Developer Approved"
+
+- **Readable**: Any developer can understand the code immediately
+- **Maintainable**: Simple patterns that are easy to modify
+- **Focused**: Each module has a single, clear responsibility
+- **Direct**: No unnecessary abstractions or indirection
+- **Practical**: Solves the specific problem without over-engineering
+
+When in doubt, choose the **simpler** solution. Your future self (and your teammates) will thank you.
+
+---
+
+## Important Instruction Reminders
+
+**For an agent when working on this project:**
+
+1. **Do what has been asked; nothing more, nothing less**
+2. **NEVER create files unless absolutely necessary for achieving the goal**
+3. **ALWAYS prefer editing an existing file to creating a new one**
+4. **NEVER proactively create documentation files (*.md) or README files**
+5. **Follow the SIMPLE patterns established in this codebase**
+6. **When in doubt, ask "Is this the simplest way?" before implementing**
+
+This project is intentionally simplified. **Respect that simplicity.**