Skip to content

TUTORIAL.md

Latest commit

 

History

History
177 lines (127 loc) · 4.89 KB

TUTORIAL.md

File metadata and controls

177 lines (127 loc) · 4.89 KB

TaskWing Tutorial

TaskWing extracts architectural knowledge from your codebase and stores it locally, giving every AI tool instant context via MCP.

Supported Models

OpenAI Anthropic Google Gemini AWS Bedrock Ollama

Works With

Claude Code OpenAI Codex Cursor GitHub Copilot Gemini CLI OpenCode

Brand names and logos are trademarks of their respective owners; usage here indicates compatibility, not endorsement.

1. Bootstrap

cd your-project
taskwing bootstrap

This creates .taskwing/ and installs AI assistant integration files.

2. Create and Activate a Plan

In your AI tool, use the MCP workflow:

/taskwing:plan

This runs clarify -> generate -> activate in one step via MCP.

3. Execute with Slash Commands

In your AI tool:

/taskwing:next

When done:

/taskwing:done

Check current status:

/taskwing:context

3.5. First-Run Success Loop (<15 minutes)

Use this minimum loop to validate TaskWing end-to-end:

  1. /taskwing:plan <goal> and approve the clarified checkpoint
  2. /taskwing:next and approve the implementation checkpoint
  3. Make a scoped change
  4. /taskwing:done with fresh verification evidence

If you complete this loop once, your setup is healthy and your assistant workflow is aligned with TaskWing contracts.

4. Inspect Progress from CLI

taskwing task list

5. MCP Server

Run MCP server when your AI tool needs stdio MCP integration:

taskwing mcp

6. Local Runtime (Optional)

Run TaskWing API/dashboard tooling locally:

taskwing start

Default bind is 127.0.0.1.

7. Troubleshooting

taskwing doctor
taskwing config show

Repair workflow:

# Apply managed repairs + MCP fixes
taskwing doctor --fix --yes

# Adopt unmanaged TaskWing-like AI files (with backup) and repair
taskwing doctor --fix --adopt-unmanaged --yes --ai claude

Bootstrap behavior during drift:

  • Managed local drift: taskwing bootstrap auto-repairs.
  • Unmanaged drift: bootstrap warns and points to doctor --fix --adopt-unmanaged.
  • Global MCP drift: bootstrap warns and points to doctor --fix.

8. Optional: Use AWS Bedrock

You can select Bedrock from the interactive config flow:

taskwing config

Or configure directly:

llm:
  provider: bedrock
  model: anthropic.claude-sonnet-4-5-20250929-v1:0
  bedrock:
    region: us-east-1
  apiKeys:
    bedrock: ${BEDROCK_API_KEY}

Recommended Bedrock model IDs:

  • anthropic.claude-opus-4-6-v1
  • anthropic.claude-sonnet-4-5-20250929-v1:0
  • amazon.nova-premier-v1:0
  • amazon.nova-pro-v1:0
  • meta.llama4-maverick-17b-instruct-v1:0

Core Commands

  • taskwing bootstrap
  • taskwing ask "<query>"
  • taskwing task
  • taskwing mcp
  • taskwing doctor
  • taskwing config
  • taskwing start

MCP Tools

Tool Description
ask Search project knowledge (decisions, patterns, constraints)
task Unified task lifecycle (next, current, start, complete)
plan Plan management (clarify, decompose, expand, generate, finalize, audit)
code Code intelligence (find, search, explain, callers, impact, simplify)
debug Diagnose issues systematically with AI-powered analysis
remember Store knowledge in project memory