eliasoulkadi/shokunin/plan
Claude Code and agent skill by eliasoulkadi/shokunin
- community
Checking that the SKILL.md file is still in the repository.
Planning agent for task breakdown and implementation planning. Use via spawn_subagent with skill='plan' when you need to explore a codebase and design an implementation approach before writing code.
Install
npx vskill@latest install eliasoulkadi/shokunin/planSKILL.md
Entered plan mode. Focus on exploring the codebase and designing an implementation approach without writing code.
In plan mode:
- Thoroughly explore the codebase to understand existing patterns
- Identify similar features and architectural approaches
- Consider multiple approaches and their trade-offs
- Design a concrete implementation strategy
- Write the plan and return its path
DO NOT write or edit any files yet (except the plan file). This is a read-only exploration and planning phase.
Workflow
Phase 1: Initial Understanding
Goal: Gain comprehensive understanding of the user's request.
Use only research agents — no other agent types.
- Understand the user's request and associated code
- Launch research agents in parallel (single message, multiple calls):
- Use
skill: 'research'for all exploration agents - Use the cheapest model available
- 1 agent for isolated tasks with known files
- Multiple agents for uncertain scope, multiple codebase areas, or pattern discovery
- Max 3 agents — usually 1 is enough
- Use multiple agents when:
- Task touches multiple parts of the codebase
- Large refactor or architectural change
- Many edge cases to consider
- Need to compare different approaches
Phase 2: Design
Goal: Design an implementation approach based on exploration results.
Launch 1-3 design agents in parallel:
- Default: 1 design agent for most tasks
- Skip agents: Only for trivial tasks (typo fixes, single-line changes, simple renames)
- Multiple agents: For complex tasks needing different perspectives
Example perspectives by task type:
- New feature: simplicity vs performance vs maintainability
- Bug fix: root cause vs workaround vs prevention
- Refactoring: minimal change vs clean architecture
For each agent:
- Provide comprehensive background from Phase 1
- Describe requirements and constraints
- Request a detailed implementation plan
Phase 3: Review
Goal: Review plans and ensure alignment with user's intentions.
- Read critical files identified by agents
- Verify plans align with the original request
- Do NOT call subagents here
Phase 4: Final Plan
Goal: Write the final plan. Include:
- Only the recommended approach, not alternatives
- Critical file paths to be modified
- Verification section (how to test end-to-end)
- Concise enough to scan, detailed enough to execute
Plan Structure Template
Common Planning Patterns
| Pattern | When to use | |
|---|---|---|
| Surgical | Small, isolated change. Read 1-2 files, write plan in 1 message. | |
| Scout | Adding to existing pattern. Find 3+ similar implementations first. | |
| Architect | New feature or refactor. Full phases 1-4 with multiple agents. | |
| Emergency | Production bug. Skip to fix, plan doc optional. |
Anti-Patterns
| Mistake | Fix | |
|---|---|---|
| Over-planning trivial changes | Small typo? Just fix it. Plan adds no value. | |
| Under-planning complex changes | New feature across 10 files? Always go through phases 1-4. | |
| Skipping verification | Always include how to test the end result. | |
| Too many alternatives in final plan | Choose one. Use the plan file, not a discussion document. | |
| Designing without reading existing patterns | Always check 3+ similar implementations first. |
Error Handling
| Cause | Fix | |
|---|---|---|
| Research subagent returns incomplete or empty results | The subagent may have hit a timeout or the codebase scope was too narrow. Re-launch with explicit file paths and a more focused query. If still empty, read the files directly. | |
| Design subagent proposes a solution that contradicts codebase conventions | Subagent lacks the full context of the codebase style guide. Manually review the plan against 3+ existing implementations. Override the subagent recommendation and document the convention in the plan. | |
| Plan file path conflicts with an existing file | The skill writes to a plan path that already exists (e.g., PLAN.md was already created). Append a timestamp: PLAN-20260518-143000.md. Ask user if they want to overwrite or keep both. | |
| Phase 1 exploration reveals the task is trivial (single-file, single-line) | Full 4-phase planning is wasteful for a typo fix. Abort plan mode. Inform user the change is trivial and ask if they want to skip directly to implementation. | |
| Plan exceeds 200 lines but user asked for concise | The plan template encourages thoroughness but the user has a scanning constraint. Collapse "Files to Modify" to a 3-column table. Collapse "Implementation Steps" to bullet points without nested details. | |
| User changes requirements mid-planning (Phase 3 or 4) | The plan was built against stale requirements. Do not patch the existing plan. Return to Phase 1 with the new requirements. Archive the old plan with a deprecation comment. |
Checklist
- [ ] Codebase explored before writing the plan — dependencies, conventions, patterns identified
- [ ] All edge cases and failure modes documented, not just happy path
- [ ] Estimations include testing, documentation, and verification time
- [ ] Dependencies and risks explicitly called out
- [ ] Plan is reviewable independently — enough context for another developer to evaluate
Sources
- "The Pragmatic Programmer" by David Thomas and Andrew Hunt (Addison-Wesley, 20th Anniversary Edition, 2019) — tracer bullet development and prototyping principles
- "Software Architecture: The Hard Parts" by Neal Ford, Mark Richards, Pramod Sadalage, Zhamak Dehghani (O'Reilly, 2021) — architectural decision records and trade-off analysis
- "A Philosophy of Software Design" by John Ousterhout (Yaknyam Press, 2nd Edition, 2021) — deep module design and complexity management
- Google Design Docs culture (google.github.io/eng-practices/review/design-docs) — structured design document template and review process
- RFC (Request for Comments) process documentation (github.com/rust-lang/rfcs) — community-driven technical proposal patterns
- "Domain-Driven Design" by Eric Evans (Addison-Wesley, 2003) — bounded context mapping and ubiquitous language in planning
- Nygard, Michael T. "Release It! Design and Deploy Production-Ready Software" (Pragmatic Bookshelf, 2nd Edition, 2018) — stability patterns and failure mode planning
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryChecking.pack/skills/plan/SKILL.md
Popularity
- Repository stars
- 112
- Repository forks
- 14
- 7-day trend score
- 24
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 27 Aug 2026, 2 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/eliasoulkadi/shokunin/plan)