agentflow-cli 0.1.0

Git-like workflow management for AI agents

AgentFlow

Git-like workflow management for AI agents.

Overview

AgentFlow is a workflow management system designed to help AI agents track and manage their work sessions in a structured, version-controlled manner similar to how Git manages code.

Features

• Workspace Management - Organize work into isolated workspaces for different projects
• Session Tracking - Track work sessions with start/end times and status
• Action Logging - Log detailed actions during active sessions
• Commit System - Create commits to summarize completed work with parent-child relationships
• Multiple Databases - Support for PostgreSQL and SQLite

Installation

pip install agentflow-cli

Or using uv:

uv pip install agentflow-cli

Quick Start

1. Initialize

Interactive setup:

agentflow init

Or with a direct database URL:

agentflow init --db-url "postgresql://user:pass@localhost/agentflow"

For SQLite:

agentflow init --db-url "sqlite:///agentflow.db"

2. Create a Workspace

agentflow workspace create my-project
agentflow workspace switch my-project

3. Start a Session

agentflow session start "Implement user authentication"

4. Log Actions

agentflow session log "Created User model"
agentflow session log "Added login endpoint"

5. Check Status

agentflow session status

Commands

Configuration

• agentflow init - Initialize configuration
• agentflow config show - Show current configuration
• agentflow config test - Test database connection

Workspace

• agentflow workspace create <name> - Create a new workspace
• agentflow workspace list - List all workspaces
• agentflow workspace switch <name> - Switch to a workspace
• agentflow workspace current - Show current workspace

Session

• agentflow session start <task> - Start a new session
• agentflow session status - Show current session status
• agentflow session abort - Abort the current session
• agentflow session log <action> - Log an action to the current session

Requirements

• Python >= 3.14
• PostgreSQL or SQLite

Development

Install development dependencies:

uv pip install -e ".[dev]"

Run tests:

uv run pytest

Lint:

uv run ruff check

Type check:

uv run mypy

License

MIT