BlogSource
How I Use Claude Code
by Boris Tane · 2026-02-11
✓ Opened with WebFetch on 2026-10-15
Read the original ↗Part 1 · What the source did
Make the agent write its research and plan to markdown files, mark up the plan yourself until it's right, and only then let it build.
- Asks Claude to read a part of the codebase deeply and write findings to research.md.
- Has it write plan.md with code snippets, file paths and trade-offs.
- Adds inline notes to the plan and sends Claude back, repeating 1–6 times with an explicit 'don't implement yet'.
- Asks for a todo list, then one standing 'implement it all' prompt that runs typecheck continuously.
“never let Claude write code until you’ve reviewed and approved a written plan”— Boris Tane, Blog
Summary written by Agent Cookbook. See the original for the full context.
Part 2 · Our copy-paste version
Adapted by us from what the source described. Fill in the [BRACKETS].
The prompt
PHASE 1: Read [FOLDER/FEATURE] in depth. Understand how it works and its edge cases. Write everything you learned to research.md. Do not change code.
PHASE 2: Based on research.md, write plan.md for [CHANGE]: approach, exact file paths, code snippets for key edits, trade-offs, open questions. Do NOT implement yet.
PHASE 3 (repeat): I added notes in plan.md. Address every note and update the plan. Do NOT implement yet.
PHASE 4: Add a granular todo list to plan.md. Do NOT implement yet.
PHASE 5: Implement the whole plan. Tick each todo in plan.md as you finish it. Run [TYPECHECK/TEST COMMAND] after each phase and fix what you broke.
🛡
Approval ruleThe agent must not touch code until you've signed off on plan.md. If it goes wrong, revert with git and narrow the scope rather than patching.
Set it up
- Start a fresh session on a clean branch.
- Run phase 1 and read research.md yourself; correct anything wrong.
- Run phase 2, then annotate plan.md in your editor.
- Repeat phase 3 until you'd approve it as a PR description.
- Run phases 4–5, review the diff, and commit.
Common failures
⚠
What usually goes wrong
- Skipping the 'don't implement yet' guard: the agent starts coding after round one.
- Thin research leads to plans that ignore existing layers. Ask for depth explicitly.
- Long sessions drift; point it back to plan.md, which survives compaction.
Variations
- Paste a reference implementation from another repo with the plan request.
- Use a second agent to review plan.md before you annotate.