dartai - Dart Task Management

dartai integrates Claude Code with Dart task management, enabling task execution, progress tracking, and dartboard management with adversarial quality verification.

Why dartai?

Seamless Dart Integration

Execute Dart tasks directly from Claude Code with automatic quality verification and documentation updates.

Key Features

• 📋 Task Execution: Execute Dart tasks through quality pipeline
• 📊 Progress Tracking: Monitor task status and dartboard activity
• 🔄 Quality Verification: Adversarial verification of implementations
• 📝 Documentation Updates: Automatic doc updates after task completion

Installation

# Install from marketplace
claude mcp add-dir https://github.com/standardbeagle/standardbeagle-tools
claude mcp add dartai --source ./plugins/dartai

Prerequisites

• Dart MCP Server: Configure dart-query MCP server
• Dart API Token: Set up authentication

Setup

# Set up Dart MCP with SLOP
/setup-dart

# Configure project roles
/setup-roles

Available Skills

Skill	Description
setup-dart	Configure dart-query MCP with SLOP
setup-roles	Configure project-specific role rules
dartai-start	Start Ralph Wiggum loop on dartboard
dartai-task	Execute single task through quality pipeline
loop-status	Show current loop status
dartai-sync	Sync local work with Dart statuses
verify	Run adversarial verification loop
dartai-config	Configure dartai settings

Usage Examples

Start Task Execution Loop

# Start loop on a dartboard
/dartai-start --dartboard "Sprint 1"

# Check status
/loop-status

Execute Single Task

# Execute a specific task
/dartai-task --task-id "TASK-123"

# With plan adjustment
/dartai-task --task-id "TASK-123" --adjust-plan

Sync with Dart

# Sync local changes to Dart
/dartai-sync

# Sync specific task
/dartai-sync --task-id "TASK-123"

Run Verification

# Run adversarial verification
/verify --type implementation

# Verify tests
/verify --type tests

# Verify security
/verify --type security

# Verify refactoring
/verify --type refactoring

Agents

task-executor

Execute Dart tasks through the adversarial quality pipeline with plan adjustment at each phase.

quality-verifier

Adversarial verification agent that challenges implementations, tests, and refactorings.

doc-updater

Update documentation after task completion including CHANGELOG, README, and Dart comments.

Hooks

dartai includes Python-based hooks for Dart integration:

Hook	Script	Purpose
SessionStart	session_init.py	Initialize session
PostToolUse (Task)	track_subagent_spawn.py	Track subagent spawns
PostToolUse (update_task)	on_task_update.py	Handle task updates
PostToolUse (list_tasks)	track_dartboard.py	Track dartboard activity
PostToolUse (Write/Edit)	track_changes.py	Track file changes
SubagentStop (task-executor)	track_iteration.py	Track task iteration
SubagentStop (quality-verifier)	track_verification.py	Track verification
SubagentStop (doc-updater)	track_docs.py	Track doc updates
Stop	session_cleanup.py	Cleanup session

Quality Pipeline

Phase 1: Task Execution

1. Load task from Dart
2. Analyze requirements
3. Implement solution
4. Write tests

Phase 2: Quality Verification

1. Independent agent reviews implementation
2. Tests verified for coverage
3. Edge cases identified

Phase 3: Documentation

1. README updated
2. CHANGELOG updated
3. Inline documentation added

Phase 4: Sync

1. Task status updated in Dart
2. Comments added for context
3. Next task identified

Configuration

Project Settings

Configure in .claude/dartai-config.yaml:

dartboard: "Sprint 1"
auto_sync: true
verification_level: "thorough"
doc_update: true

roles:
  - name: backend-developer
    triggers:
      - file_pattern: "src/api/**"
    actions:
      - run_tests: true
      - security_audit: true

Environment Variables

# Dart API token
export DART_API_TOKEN="your-token"

# Default dartboard
export DARTAI_DEFAULT_DARTBOARD="Sprint 1"

MCP Integration

dartai uses the dart-query MCP server:

{
  "mcpServers": {
    "dart-query": {
      "command": "npx",
      "args": ["-y", "dart-query-mcp@latest"]
    }
  }
}

Troubleshooting

Dart Connection Failed

# Verify API token
echo $DART_API_TOKEN

# Test connection
dart-query test-connection

Tasks Not Syncing

1. Check auto_sync is enabled
2. Verify dartboard name matches
3. Ensure MCP server is running

Verification Failing

# Run verification with verbose output
/verify --type implementation --verbose

Related Resources

• workflow Plugin - General task automation
• SLOP MCP - MCP server management
• Dart Documentation

Version History

Version	Changes
0.1.0	Initial release with quality pipeline