Lark Agent Skill

version: 1.0.0 author: Claude Code tags: [lark, testing, automation, markdown, json, task-management]

Lark Agent Skill

Create structured Lark tasks from markdown test documentation with proper hierarchy and verification.

Overview

The Lark Agent skill provides a complete end-to-end workflow for:

1. Parsing markdown test files with clear structure
2. Generating structured JSON with test hierarchy
3. Preparing Lark task creation workflow (3-level hierarchy)
4. Preparing verification workflow
5. Outputting complete workflow plan for execution

This skill is designed for teams that maintain test documentation in markdown format and need to convert them into trackable Lark tasks automatically.

When to Use This Skill

Activate this skill when user:

• Wants to create Lark tasks from test plans
• Mentions "lark", "test cases", "test planning", "test scenarios"
• Has a markdown file with test documentation
• Needs to convert test documentation into Lark tasks
• Wants to process test files into structured tasks

Workflow

1. Markdown Input Processing

The skill accepts markdown files with the following structure:

# Test Title
Test description

## Test Scenario: Scenario Name
Scenario description

### Task: Task Name
1. Step one
2. Step two
Expected Result: What should happen

2. JSON Generation

Converts markdown into structured JSON:

{
  "testOverview": {
    "title": "Test Title",
    "description": "Test description",
    "owner": "Assigned User",
    "targetDate": "2025-11-02",
    "status": "pending"
  },
  "scenarios": [
    {
      "scenarioId": "scenario-0-timestamp",
      "title": "Scenario Name",
      "description": "Scenario description",
      "tasks": [
        {
          "taskId": "task-0-0-timestamp",
          "title": "Task Name",
          "description": "Steps",
          "expectedResult": "Expected outcome",
          "status": "pending"
        }
      ]
    }
  ]
}

3. Lark Task Creation

Creates hierarchical tasks in Lark via MCP:

• Level 1: Parent task (test overview)
• Level 2: Scenario tasks (marked as milestones)
• Level 3: Individual test tasks

Uses Lark MCP tools:

• task_v2_tasklist_create - Create task list
• task_v2_task_create - Create parent task
• task_v2_taskSubtask_create - Create scenario and individual tasks
• task_v2_task_addMembers - Assign users

4. Verification & Reporting

After task creation, the skill verifies:

• ✅ Parent task created with correct details
• ✅ All scenario tasks created as milestones
• ✅ All individual tasks created with test steps
• ✅ Task hierarchy is correct (parent → scenarios → tasks)
• ✅ User assignments are correct
• ✅ Dates and priorities match specifications

Generates a comprehensive report showing:

• Number of tasks created at each level
• Any failures or issues encountered
• Confirmation that structure matches requirements
• Recommendations for any issues found

Usage

⚠️ Important: Interactive Mode Not Supported

Interactive mode (stdin prompts) does NOT work in Claude Code's environment.

The skill works via conversational approach with visual indicators.

🎯 How It Works

When user requests Lark task creation:

1️⃣ Show Activation Banner

Display:

╔══════════════════════════════════════════════════════════════════╗
║              🚀 LARK AGENT SKILL ACTIVATED 🚀                   ║
║      Converting Test Plans → Structured Lark Tasks              ║
╚══════════════════════════════════════════════════════════════════╝

2️⃣ Collect Parameters with Emojis

• 📄 Test file path (required): "📄 Which test file would you like to process?"
• 👤 Owner name (optional): "👤 Who should be the task owner? (default: QA Team)"
• ⚡ Priority (optional): "⚡ What priority? (1=low, 2=medium, 3=high, default: 2)"
• 📅 Target date (optional): "📅 What's the target completion date? (YYYY-MM-DD)"

3️⃣ Execute Skill with Progress Indicators

Show:

🔄 Executing Lark Agent skill...
📂 File: [file-path]
👤 Owner: [owner]
⚡ Priority: [priority]
📅 Target: [date]

Execute:

python .claude/​skills/​lark-agent/​run.py [file-path] \
  --owner="[name]" \
  --priority=[1-3] \
  --target-date="YYYY-MM-DD"

4️⃣ Parse Output

Show: 📊 Parsing workflow output...

5️⃣ Execute Lark MCP Workflow

Show progress:

🏗️ Creating Lark tasks...
   ✅ Step 1: Creating task list
   ✅ Step 2: Creating parent task
   ✅ Step 3: Creating scenario tasks
   ✅ Step 4: Creating individual tasks

6️⃣ Report Results

Show completion banner:

╔══════════════════════════════════════════════════════════════════╗
║              ✅ LARK AGENT WORKFLOW COMPLETED! ✅               ║
╚══════════════════════════════════════════════════════════════════╝

📊 Summary:
   📋 Test: [test title]
   🎯 Scenarios: [count]
   📝 Total Tasks: [count]

Direct Mode (Only Working Mode)

python .claude/​skills/​lark-agent/​run.py [file-path] --owner="name" --priority=2 --target-date="YYYY-MM-DD"

Available Options

• file-path (required): Path to markdown test file
• --owner (optional): Task owner name (default: "QA Team")
• --priority (optional): 1=low, 2=medium, 3=high (default: 2)
• --target-date (optional): Target date YYYY-MM-DD (default: 7 days from today)
• --task-list-id (optional): Existing Lark task list ID

Example Conversation

User: Create Lark tasks from tests/​manual/​login-test.md
Claude: I'll help you create Lark tasks. Who should be the owner?
User: QA Team
Claude: What priority? (1=low, 2=medium, 3=high)
User: 2
Claude: What's the target date? (YYYY-MM-DD)
User: 2025-12-31
Claude: [Executes skill and creates tasks]

Implementation Details

Scripts

The skill uses the following Python scripts located in scripts/:

1. lark_agent.py - Main entry point and workflow orchestrator
2. markdown_parser.py - Parses markdown and extracts test structure
3. lark_task_creator.py - Creates hierarchical Lark tasks via MCP
4. lark_task_verifier.py - Verifies task creation and generates reports

These scripts are designed to work with Claude Code's tool calling capabilities for Lark MCP integration.

References

Documentation in references/:

1. usage-guide.md - Detailed usage instructions and examples
2. json-schema.md - Complete JSON structure specification
3. markdown-format.md - Markdown format requirements

Templates

Example files in assets/​templates/:

1. test-template.md - Template for creating new test files
2. output-template.json - Example JSON output structure

Integration with Lark MCP

This skill uses the following Lark MCP tools:

• task_v2_task_create - Create tasks
• task_v2_tasklist_create - Create task lists
• task_v2_taskSubtask_create - Create subtasks
• task_v2_task_addMembers - Assign users
• timezone tools - Handle date calculations

Error Handling

The skill handles common errors:

• File not found: Validates file path before processing
• Invalid markdown structure: Reports parsing errors with line numbers
• Lark API errors: Retries failed task creation and reports issues
• Date validation: Ensures dates are valid and in the future

Best Practices

Markdown File Structure

• Use clear heading hierarchy (H1 for title, H2 for scenarios, H3 for tasks)
• Include expected results for each task
• Keep task descriptions concise but complete
• Use consistent naming conventions

Task Organization

• Group related tests into scenarios
• Set realistic target dates
• Assign appropriate owners
• Use meaningful task titles

Examples

See references/​usage-guide.md for detailed examples and assets/​templates/ for template files.

Troubleshooting

Markdown Parsing Issues

If markdown parsing fails:

1. Check heading hierarchy (H1 > H2 > H3)
2. Ensure proper markdown syntax
3. Verify file encoding (UTF-8)

Lark Task Creation Issues

If task creation fails:

1. Verify Lark MCP server is running
2. Check user permissions in Lark
3. Validate date formats
4. Review error messages in output

Version History

v1.0.0

• Initial release
• Markdown parsing and JSON generation
• Hierarchical Lark task creation
• Basic error handling and validation