This skill should be used when the user asks to "build", "implement", "create", "review", "audit", "debug", "fix", "plan", or "design" something. Routes requests to appropriate agents (component-builder, code-reviewer, bug-investigator, planner) based on detected intent.
This skill inherits all available tools. When active, it can use any tool Claude has access to.
Route user requests to the appropriate agent based on intent detection.
Detect intent from user request and route to the appropriate workflow:
| Intent Keywords | Workflow | Agent Chain |
|---|---|---|
| build, implement, create, write, add, make | BUILD | component-builder → code-reviewer → integration-verifier |
| review, audit, check, analyze, assess | REVIEW | code-reviewer |
| debug, fix, error, bug, troubleshoot, broken | DEBUG | bug-investigator → code-reviewer → integration-verifier |
| plan, design, architect, roadmap, strategy | PLAN | planner |
When user wants to build/implement/create something:
When user wants to review/audit code:
When user encounters errors/bugs:
When user wants to plan/design:
Use Task tool to invoke agents:
Task(subagent_type="cc10x:component-builder", prompt="Build [component description]")
Task(subagent_type="cc10x:code-reviewer", prompt="Review [code/PR description]")
Task(subagent_type="cc10x:bug-investigator", prompt="Debug [error description]")
Task(subagent_type="cc10x:integration-verifier", prompt="Verify [integration description]")
Task(subagent_type="cc10x:planner", prompt="Plan [feature description]")
Always run agents sequentially when outputs depend on previous steps.
After workflow completion, provide:
## Workflow Summary
- **Intent**: [BUILD/REVIEW/DEBUG/PLAN]
- **Agents Used**: [list]
- **Result**: [success/issues found]
## Key Outcomes
[Summary of what was accomplished]
## Next Steps
[If any follow-up needed]