Discipline

The Lab Notes Practice Playbook

How to keep an agent productive across sessions instead of relearning the repo.

LAB-NOTES-PRACTICE-PLAYBOOK.md · 12 sections · 1,579 words · 11 KB · markdown

The meta-playbook. How to maintain a CLAUDE.md that stays accurate, when to append a lab note, what belongs in project memory versus the repo, and the debugging rule that stops an agent from iterating the same failed fix three times.

What's inside

//The CLAUDE.md structure that survives multi-session work
//Lab-note format: date · symptom · attempted fix · why it failed · resolution
//What belongs in memory vs the repo vs a lab note
//Hygiene rules so the file doesn't rot into noise
//The read-the-logs-first rule before any second attempt

When to use it

Continuously. Every session.

Build these first

Nothing. This one stands alone.

Straight from the file

The CLAUDE.md file at the root of each app repo is the canonical agent-instructions doc. It accumulates four kinds of knowledge that the agent must read before doing anything:

Table of contents

0. What this is
1. Why it matters
2. Lab notes format (strict)
3. When to append a lab note
4. CLAUDE.md section structure (mirror BKE's)
5. Hygiene rules
6. The memory layer (complement, not substitute)
7. Verify before citing
8. Bootstrapping CLAUDE.md for a new app
9. Agent kickoff prompt
10. Common pitfalls
11. Worked example

How to run it

Drop LAB-NOTES-PRACTICE-PLAYBOOK.md into your repo, then tell Claude Code: “Read LAB-NOTES-PRACTICE-PLAYBOOK.md and implement it in this codebase.” It's written for an agent to execute end to end. Search the file for angle-bracket placeholders and swap in your own project names, domains, and keys before the first run.

↓ Download the playbook

More discipline playbooks

C-tier

← All 28 Claude templates

The Lab Notes Practice Playbook

Pre-Production Hardening

Supabase Migrations

Core Web Vitals Audit