EXECUTE IMMEDIATELY

Step 1: Validate Input

If args is empty or only whitespace:
  Ask user: "What would you like me to implement?"
  STOP and wait for user response

If args length > 2000 characters:
  Truncate to 2000 characters
  Warn user: "Request was truncated to 2000 characters"

Step 2: Domain Detection & Routing

Read ${CLAUDE_PLUGIN_ROOT}/agents/_shared/domain-router.json to determine the target agent.

Detection logic (apply in order):

1. Extract file extensions from args (e.g., "fix Login.tsx" → .tsx → frontend)
2. Check args text against domain keywords in domain-router.json
3. Check for path patterns (e.g., ".github/workflows/" → devops)
4. Check for known filenames (e.g., "Dockerfile" → devops)

Routing decision:

• Frontend match → subagent_type="team-shinchan:aichan"
• Backend match → subagent_type="team-shinchan:bunta"
• DevOps match → subagent_type="team-shinchan:masao"
• No match → subagent_type="team-shinchan:bo" (fallback)

Step 3: Execute Task

Do not read further. Execute this Task NOW:

Task(
  subagent_type="{detected_agent from Step 2}",
  model="sonnet",
  prompt=`/team-shinchan:implement has been invoked.

## Implementation Request

Handle coding tasks including:

| Area | Capabilities |
|------|-------------|
| Feature Implementation | New features, functions, classes |
| Bug Fixes | Debugging, error correction |
| Code Modification | Refactoring, updates, changes |
| Utilities | Helper functions, utilities |
| Tests | Unit tests, integration tests |

## Implementation Requirements

- Read existing code first to understand patterns
- Follow project conventions
- Write clean, maintainable code
- Handle errors gracefully
- Keep functions small and focused
- Add comments only for complex logic

## Post-Implementation Verification

After writing code:
1. Run existing tests if available (detect test framework from package.json/config)
2. If tests fail, fix the issues before reporting completion
3. If no tests exist, verify the code compiles/loads without errors

## Output Format

After implementation:
- Summary of changes made
- Files modified with line references
- Test results (pass/fail/skipped)
- Any follow-up recommendations

User request: ${args || '(Please describe what to implement)'}
`
)

STOP HERE. The above Task handles everything.