deepsecure: Instructions file for Claude Code

CLAUDE.md

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

Project Overview

DeepSecure is a security platform that provides Identity-as-Code for AI agents, enabling them to fetch their own ephemeral credentials programmatically instead of using static API keys. The project consists of a Python CLI/SDK and backend services that implement a dual-service gateway architecture.

Development Commands

Environment Setup

# Install development dependencies
make install-dev
# or traditionally
make install-traditional

# Setup development environment
make setup

Testing

# Run all tests
make test
pytest

# Run tests with coverage
make test-cov
pytest --cov=deepsecure --cov-report=html --cov-report=term

# Run specific test markers
pytest -m e2e          # End-to-end tests (require live backend)
pytest -m integration  # Integration tests

Code Quality

# Run linting
make lint
ruff check .
mypy deepsecure/

# Format code
make format
black .
isort .

# Security scanning
make security
bandit -r deepsecure/
safety check

# Run all quality checks
make check-all

Build and Package

# Build package
make build
./scripts/build_package.sh

# Clean build artifacts
make clean

Architecture Overview

Core Module Structure

• deepsecure/_core/: Internal implementation modules

  • client.py: High-level client services wrapping core functionalities
  • base_client.py: Base HTTP client with authentication
  • vault_client.py: Secret management operations
  • agent_client.py: Agent identity management
  • identity_manager.py: Agent identity and key management
  • crypto/key_manager.py: Cryptographic key operations
  • config.py: Configuration management

• deepsecure/: Public API layer

  • client.py: Main SDK client (uses _core modules)
  • commands/: CLI command implementations
  • integrations/: Framework integrations (LangChain, CrewAI)
  • resources/: Resource objects (Agent, Secret)