
implementation-planning
by bnadlerjr
Various dotfiles
SKILL.md
name: implementation-planning description: Create detailed implementation plans through interactive research. Use when in plan mode or when creating technical specifications.
Implementation Planning
Create detailed implementation plans through an interactive, iterative process. Be skeptical, thorough, and collaborative.
When This Skill Applies
- You're in plan mode (
permission_mode: plan) - User asks to create an implementation plan
- User references a ticket/task to plan
Core Principles
- Interactive: Never dump a complete plan. Gather context → verify understanding → align on approach → detail phases.
- Grounded: Every claim verified against actual code. Include
file:linereferences. - Bounded: Every plan MUST include "What We're NOT Doing" section.
- Testable: Separate automated vs manual success criteria.
Process
Step 1: Context Gathering
-
Read all mentioned files FULLY (tickets, research, existing plans)
- Use
$(claude-docs-path tickets)/for tickets - Use
$(claude-docs-path research)/for research docs - Use
$(claude-docs-path plans)/for existing plans - NEVER read files partially
- Use
-
Spawn research agents in parallel:
- codebase-locator → Find all relevant files
- codebase-analyzer → Understand current implementation
- docs-locator → Find existing documentation
- jira-cli-expert → Get ticket details (if Jira mentioned)
-
Present informed understanding:
Based on the ticket and my research, I understand we need to [summary]. I've found: - [Current implementation detail with file:line] - [Relevant pattern or constraint] - [Potential complexity identified] Questions my research couldn't answer: - [Question requiring human judgment]Only ask questions you genuinely cannot answer through code investigation.
Step 2: Research & Discovery
After clarifications:
-
Verify corrections: If user corrects you, spawn new research tasks. Don't just accept—verify.
-
Track with TodoWrite: Create a research checklist
-
Present findings and options:
Based on my research: **Current State:** - [Key discovery with file reference] - [Pattern to follow] **Design Options:** 1. [Option A] - [pros/cons] 2. [Option B] - [pros/cons] **Open Questions:** - [Technical uncertainty] - [Design decision needed] Which approach aligns best?
Step 3: Structure First
Before detailing, present outline:
Proposed plan structure:
## Overview
[1-2 sentences]
## Phases:
1. [Phase name] - [what it accomplishes]
2. [Phase name] - [what it accomplishes]
Does this phasing make sense?
Get feedback before proceeding.
Step 4: Detailed Plan
When user approves structure, present the complete plan. Use the template in templates/plan-template.md.
Important: In plan mode, when you exit via ExitPlanMode, the plan will be automatically:
- Saved to
$(claude-docs-path plans)/ - Given proper Obsidian frontmatter
- Named based on the H1 header
So ensure your plan has a clear H1 header like # [Feature Name] Implementation Plan.
Guidelines
Do
- Read files FULLY before planning
- Include file:line references for all claims
- Use custom agents for research (they work in plan mode)
- Verify claims against code
- Get buy-in at each step
- Separate automated vs manual verification
Don't
- Write complete plans before alignment
- Accept corrections without verification
- Leave open questions in final plan
- Assume—verify with code
No Open Questions Rule
If you encounter open questions during planning:
- STOP
- Research or ask for clarification immediately
- Do NOT present a plan with unresolved questions
- Every decision must be made before finalizing
Agent Reference
Use these agents for research (all read-only, work in plan mode):
| Agent | Purpose |
|---|---|
codebase-locator | Find relevant files |
codebase-analyzer | Understand implementation details |
codebase-pattern-finder | Find similar features to model after |
docs-locator | Find existing documentation |
docs-analyzer | Extract insights from documents |
jira-cli-expert | Jira operations (search, create, transition, etc.) |
Common Patterns
Database Changes
- Schema/migration
- Store methods
- Business logic
- API endpoints
- Client updates
New Features
- Research existing patterns
- Data model
- Backend logic
- API layer
- UI last
Refactoring
- Document current behavior
- Incremental changes
- Backwards compatibility
- Migration strategy
Score
Total Score
Based on repository quality metrics
SKILL.mdファイルが含まれている
ライセンスが設定されている
100文字以上の説明がある
GitHub Stars 100以上
3ヶ月以内に更新がある
10回以上フォークされている
オープンIssueが50未満
プログラミング言語が設定されている
1つ以上のタグが設定されている
Reviews
Reviews coming soon