plan-implementation

Create implementation plans for features and changes. Use when user says "plan how to implement user authentication", "how should I add dark mode to the app", "what's the best way to refactor the database layer", "create a plan for migrating to the new API", or "I need to implement [feature] - help me plan it out".

Stars 30

Forks 9

Install this agent skill to your Project

npx add-skill https://github.com/Uniswap/ai-toolkit/tree/next/packages/plugins/development-planning/skills/plan-implementation

SKILL.md

Implementation Planner

Create clear, actionable implementation plans through collaborative multi-agent discussion and refinement.

When to Activate

User wants to implement a new feature
User asks "how should I implement X?"
User wants to add functionality
User needs to refactor or migrate code
Complex bug fixes requiring planning
Architectural changes or design decisions
Any task that benefits from structured planning

Quick Process

Analyze Context: Understand the task, leverage any prior exploration
Select Agents: Choose 3-10 specialists based on task complexity
Collaborative Discussion: 2-3 rounds of multi-agent refinement
Synthesize Plan: Generate comprehensive implementation plan
Output File: Write plan to .claude-output/plan-[timestamp].md

Complexity-Based Planning

Task Type	Agents	Rounds	Plan Length
Simple (bug fix, minor feature)	3-4	1-2	~100-200 lines
Medium (features, refactors)	5-7	2-3	~200-400 lines
Complex (architecture, major features)	8-10	2-3	~400-600 lines

Plan Structure

Generated plans include:

Overview - High-level summary
Scope - What will/won't be implemented
Current State - Relevant architecture and files
API Design (optional) - Interfaces and data structures
Implementation Steps - Clear sequential steps (5-7 typical)
Files Summary - Files to create/modify
Critical Challenges (optional) - Risks and mitigations
Agent Collaboration Summary - Consensus and trade-offs

Workflow Integration

This is Step 2 of the implementation workflow:

Explore → 2. Plan (this) → 3. Review → 4. Execute

Automatically uses context from prior exploration
Output plan can be reviewed with /review-plan
Approved plans execute with execute-plan skill

Detailed Reference

For agent selection, discussion protocols, and output formats, see planning-guide.md.