AI Coding Agents & Rules
Target: Claude Code, Cursor, Windsurf, Cline, Roo Code

Subagent Task Delegation & Verification Prompt

Prompt orchestrating agents to delegate subtasks to child agents with strict verification gates and bounded context.

SubagentsOrchestrationAI AgentsClaude CodeAutomation

Interactive Prompt Playground

Customize Template Fields
Live replacement
{{SUBTASK_OBJECTIVE}}
{{ALLOWED_FILES}}
{{READONLY_FILES}}
{{VERIFICATION_COMMAND}}
Rendered Prompt (Ready to paste)1318 characters
You are a Lead Orchestrator Agent delegating an isolated subtask to a specialized Child Subagent.

Delegate the following subtask with strict context boundaries and acceptance criteria:

Subtask Objective:
```
Implement the Base64 encode/decode utility component with UTF-8 support, error handling for malformed input, and copy-to-clipboard button.
```

File Scope & Boundaries:
- Allowed Files to Modify: components/dev-tools/base64-page.tsx, lib/dev-tools/base64.ts
- Read-Only Context Files: components/ui/button.tsx, lib/utils.ts, components/site-header.tsx

Delegation Instructions to Child Agent:
1. **Strict Context Isolation**:
   - You may ONLY modify the files listed under Allowed Files. Do not edit external files.
   - Reuse existing utilities in `components/ui/button.tsx, lib/utils.ts, components/site-header.tsx` rather than creating duplicates.
2. **Execution & Code Quality**:
   - Follow strict TypeScript typing with zero `any`.
   - Adhere to the existing repository code style and design tokens.
3. **Self-Verification Protocol**:
   - Run `npx tsc --noEmit` and ensure zero errors before returning.
4. **Structured Completion Report**:
   - Return a concise report containing:
     - Changes made per file
     - Verification command output
     - Any follow-up notes for the parent orchestrator

How to Use This Prompt

  1. Define the specific subtask, allowed files, and verification command.
  2. Send to your orchestrator or subagent runner.
  3. Receive cleanly isolated, verified code changes.

Engineering Tips & Best Practices

  • Never give a subagent full repository write access if it only needs to touch 1-2 files.

What This Prompt Inspects

Key failure modes, design principles, and quality standards evaluated during execution.

Strict File Boundary

Restricts subagent edits to prevent unintended modifications to other modules.

Verification Gate

Demands child agent run compiler and test checks before reporting completion.

Structured Return Report

Ensures parent orchestrator receives clean summary of changes.

SprintKit Workflow Integrations

Complementary interactive tools and workflows across SprintKit to accelerate your engineering process.

Related Prompts

Explore related developer prompts in this workflow domain.

All Prompts

Autonomous Coding Agent Implementation Plan Generator

Instruct coding agents (Cursor, Claude Code, Cline) to generate actionable multi-step implementation plans before touching code.

Claude CodeCursorClineAI Agents+1
2 variablesCustomize prompt

Cursor & Windsurf Rules Generator (.cursorrules / .windsurfrules)

Generate a custom, context-rich project rules file that stops AI coding agents from making recurring mistakes.

CursorWindsurfAI AgentsCoding Agents+1
3 variablesCustomize prompt

Test-Driven Development (TDD) Agent Loop Prompt

Force AI coding agents to follow the Red-Green-Refactor cycle: write failing tests first, then write minimal code to pass.

TDDTestingAI AgentsCursor+1
2 variablesCustomize prompt