CLEO

About

CLEO (Command Line Entity Orchestrator) is a vendor-neutral brain and memory system specifically designed for AI software development. It establishes a robust contract between developers and their AI coding agents, providing a structured protocol for portable memory, immutable provenance, and deterministic execution. Unlike traditional task managers, CLEO prioritizes agent needs, validating every operation to prevent hallucinations, persisting state across sessions to overcome context loss, and outputting structured JSON for clarity. It's optimized for tools like Claude Code but is intentionally portable across various model providers, agent runtimes, and coding environments, making it ideal for managing complex projects with AI assistance.

Key Features

• Interoperable Interfaces: Provides a command-line interface (CLI) and MCP (Managed Code Protocol) server with provider/tool adapters to avoid vendor lock-in across AI tools.
• LLM-Agent-First Design: Defaults to structured JSON output and uses precise exit codes, with human-readable formatting available via a `--human` flag.
• Anti-Hallucination Protection: Employs four layers of validation (Schema, Semantic, Cross-File, State Machine) to prevent AI-generated errors and ensure data integrity.
• Portable Memory & Provenance: Organizes tasks in a Project -> Epic -> Task hierarchy, maintaining immutable audit trails and tracing every artifact to its origin.
• Stable Task IDs: Uses flat, sequential, and eternal IDs (e.g., T001) that decouple identity from hierarchy, ensuring external references remain valid forever.
• 104 GitHub stars

Use Cases

• Managing and tracking development tasks for AI coding agents like Claude Code.
• Providing persistent, validated memory and context for AI agents across development sessions.
• Ensuring deterministic and auditable execution of AI-generated operations in software projects.