Quick Reference: Four High-Leverage Claude Code Cards
Recall four compact decision rules for choosing Claude Code lanes and evidence paths.
When should you spawn a subagent? When the next step is noisy search with low retention value for the main thread. Good fits: log triage, wide repo scans, and evidence gathering where the lead session mainly needs ranked findings back. Which workflow fits? Known lanes vs unknown terrain Use predictability as the deciding axis, not just how big the task feels. Hero prompt temptation "I will just ask Claude Code to do the whole migration in one shot." Start with inspect -> change -> verify instead. One-shot ambition usually hides missing evidence and missing stop rules. It reframes speed around…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in