# Sprint Plan

> MUST use when user describes a multi-step feature to build, says 'let's build', 'new project', 'let's implement', or when a task clearly requires more than 3 steps and no plan exists yet. Do NOT trigger for simple additions like adding a comment, renaming a variable, or single-file changes.

- Skill: `lethilu4796/sprint-plan` (Agent Skill)
- Install (CLI): `npx skillmds@latest add lethilu4796/sprint-plan`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lethilu4796/sprint-plan/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: lethilu4796 (https://skillmd.com/u/lethilu4796)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/lethilu4796/sprint-plan

---


Create a sprint plan for: $ARGUMENTS

1. **Analyze the requirement**: Understand scope, constraints, dependencies
2. **Check existing code**: Search codebase for related implementations that can be reused
3. **Break into tasks**:
   - Each task should be independently deliverable
   - Include file paths that will be modified or created
   - Estimate size: S (< 1hr), M (1-4hr), L (4-8hr), XL (> 1 day)
4. **Sequence tasks**: Identify dependencies, mark blocking tasks, find parallelizable work
5. **Identify risks**: What could go wrong? What needs clarification?
6. **Output format**:

## Sprint Plan: [Feature Name]

### Tasks
| # | Task | Size | Depends On | Files |
|---|------|------|-----------|-------|
| 1 | ... | S | - | path/to/file |

### Risks
- [Risk description and mitigation]

### Definition of Done
- [ ] All tasks completed
- [ ] Tests written and passing
- [ ] Code reviewed
- [ ] Documentation updated
- [ ] No security vulnerabilities introduced