/plan
Activate deep planning mode for: $ARGUMENTS
$ npx -y skills add vignesh2027/claude-best-practice --agent claude-codeHow it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/plan
Context preview
What this command does when you run it.
Activate deep planning mode for: $ARGUMENTS
Command definition
plan.md/plan — Enter Structured Planning Mode
Activate deep planning mode for: $ARGUMENTS
Instructions
Before writing a single line of code or making any file change, produce a complete implementation plan using the following structure:
---
1. Goal Restatement
Restate the goal in your own words to confirm understanding.
2. Affected Files
List every file that will be:
- **Created** (new files)
- **Modified** (existing files changed)
- **Deleted** (files to be removed)
- **Left alone** (explicitly confirm nothing else is touched)
3. Implementation Steps
Number each step. Be specific — not "update auth" but "add `refreshToken` field to `UserSession` type in `src/types/auth.ts:42`".
4. Test Strategy
- What existing tests need to be updated?
- What new tests should be written?
- What's the command to run them?
5. Risks & Irreversible Steps
Highlight any:
- Database migrations
- Schema changes
- Breaking API changes
- Secrets/credential changes
6. Rollback Plan
If something goes wrong at step N, how do we restore the previous state?
7. Checkpoints
After which steps should I review the diff before continuing?
---
**DO NOT EXECUTE until the plan is reviewed and approved.**
The most comprehensive, production-ready guide to mastering Claude Code From vibe coding → agentic engineering → autonomous AI development teams Built for engineers who treat Claude Code as infrastructure, not a toy.
Repo: vignesh2027/claude-best-practice

