# TaskGraph - [TaskGraph: task management for Claude Code](https://taskgraph-ai.com/introduction.md): TaskGraph turns a PRD into a dependency-aware task list for Claude Code, then works it one task at a time through slash commands, a CLI or an MCP server. - [Upgrading](https://taskgraph-ai.com/upgrading.md): What to change when moving an existing project to a newer TaskGraph, from Task Master or from any earlier release - [Migrating from Task Master](https://taskgraph-ai.com/migrating-from-taskmaster.md): Import a Task Master project, with every tag, its docs and reports, into TaskGraph with one command - [Quick Start](https://taskgraph-ai.com/getting-started/quick-start/quick-start.md): Get started with TaskGraph in minutes: install it, choose Claude models, parse your first PRD into tasks and start working them in Claude Code. - [Requirements](https://taskgraph-ai.com/getting-started/quick-start/requirements.md): What TaskGraph needs: Claude Code installed and signed in, and Node.js 22.13 or later. No API keys or provider accounts. - [Installation](https://taskgraph-ai.com/getting-started/quick-start/installation.md): Install the taskgraph-ai npm package, run taskgraph init in your project, and add the Claude Code plugin for /taskgraph: slash commands. - [Configuration](https://taskgraph-ai.com/getting-started/quick-start/configuration-quick.md): Choose which Claude models TaskGraph uses for its main, research and fallback roles. No API keys needed. - [PRD Creation and Parsing](https://taskgraph-ai.com/getting-started/quick-start/prd-quick.md): Write a PRD (product requirements document) and parse it into a dependency-aware TaskGraph task list with Claude Code. - [Tasks Setup](https://taskgraph-ai.com/getting-started/quick-start/tasks-quick.md): Review a parsed task plan: expand tasks into subtasks, check dependencies and complexity, and get tasks ready for your coding agent. - [Executing Tasks](https://taskgraph-ai.com/getting-started/quick-start/execute-quick.md): Work through your TaskGraph tasks in Claude Code: find the next task, discuss the plan, implement it, log progress and mark it done. - [Authentication](https://taskgraph-ai.com/getting-started/api-keys.md): TaskGraph runs on your signed-in Claude Code. No provider API keys are needed. - [Configuration](https://taskgraph-ai.com/configuration.md): Configure TaskGraph models, defaults and environment variables - [Tutorial](https://taskgraph-ai.com/tutorial.md): Set up TaskGraph and use it for AI-driven development in Claude Code - [Example Interactions](https://taskgraph-ai.com/examples.md): Common Claude Code interactions when using TaskGraph - [FAQ](https://taskgraph-ai.com/getting-started/faq.md): Answers to common TaskGraph questions and where to get help with Claude Code task management. - [Contributing to TaskGraph](https://taskgraph-ai.com/getting-started/contribute.md): How to contribute to TaskGraph: set up the repo, run the tests, and open a pull request with a conventional-commit title. - [Intro to Advanced Usage](https://taskgraph-ai.com/best-practices/index.md): Advanced tips, recommended workflows and best practices for getting the most out of TaskGraph in Claude Code. - [Advanced Configuration](https://taskgraph-ai.com/best-practices/configuration-advanced.md): Advanced TaskGraph settings: which MCP tools load, timeouts, tagged task lists for branches and features, and the state file. - [Advanced Tasks](https://taskgraph-ai.com/best-practices/advanced-tasks.md): The day-to-day TaskGraph workflow in Claude Code: pick the next task, implement it, log notes, update subtasks and mark work done. - [MCP Tools](https://taskgraph-ai.com/capabilities/mcp.md): The TaskGraph MCP server: how its Model Context Protocol tools expose task management to Claude Code and other MCP clients. - [Claude Code Integration](https://taskgraph-ai.com/capabilities/claude-code.md): How TaskGraph uses Claude Code for every AI operation, and the advanced settings it exposes - [CLI Commands](https://taskgraph-ai.com/capabilities/cli-root-commands.md): A reference of all TaskGraph commands, grouped by what you use them for - [Finding Work](https://taskgraph-ai.com/capabilities/cli-viewing-commands.md): List tasks, pick the next one, see what blocks a task, and prime an agent - [Updating Tasks](https://taskgraph-ai.com/capabilities/cli-updating-commands.md): Change tasks, log notes, set status, manage dependencies and move tasks - [Planning Commands](https://taskgraph-ai.com/capabilities/cli-planning-commands.md): Turn a PRD into tasks, add and break down tasks, size them, and research - [Project Setup](https://taskgraph-ai.com/capabilities/cli-project-commands.md): Initialize a project, manage tags and models, set up tab completion, and run the loop - [Task Structure](https://taskgraph-ai.com/capabilities/task-structure.md): Tasks in TaskGraph follow a specific format designed to provide comprehensive information for both humans and AI assistants. - [Dependency Graph](https://taskgraph-ai.com/capabilities/dependency-graph.md): How TaskGraph picks the next task, finds bottlenecks, the critical path and parallel waves, explains blocked tasks, and records where each dependency came from - [Archiving finished tasks](https://taskgraph-ai.com/capabilities/archive.md): Move done and cancelled tasks out of tasks.json, keep their IDs reserved, and restore one by ID - [Loop Command](https://taskgraph-ai.com/capabilities/loop.md): Run Claude Code in autonomous loops for batch task completion - [Run Tasks in Parallel](https://taskgraph-ai.com/capabilities/loop-parallel.md): Run independent tasks side by side, each in its own git worktree, with loop --parallel - [Research](https://taskgraph-ai.com/capabilities/research.md): Web research from Claude Code with no extra API keys, plus optional Context7 library docs - [GitHub Sync](https://taskgraph-ai.com/capabilities/github-sync.md): Push your tasks to GitHub Issues from the CLI or an MCP client, with a preview first, and pull issue status back - [Merging tasks.json](https://taskgraph-ai.com/capabilities/merge-driver.md): A Git merge driver that keeps both branches' new and archived tasks, renumbers the incoming ones, and marks real clashes as conflicts - [JSON Output](https://taskgraph-ai.com/capabilities/json-output.md): Which commands print JSON, the schemaVersion every payload carries, and what stays stable within a version ## Optional - [Github](https://github.com/stephennmiller/taskgraph-ai) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.