5 hacks
Write the plan in one session, then open a fresh Claude with no context and ask it to review the plan like a staff engineer.
End implementation prompts with /simplify so parallel agents review reuse, efficiency, quality, and CLAUDE.md compliance in one sweep.
Force verification with prompts like “prove this works” and “grill me before the PR” so Claude has to validate instead of declaring done.
When a skill hits an edge case, append it to a gotchas section in that skill rather than to CLAUDE.md. The lesson then loads only when the skill runs.
Auto-compact waits until your context is nearly full, which is well past the point where answer quality starts sliding. Set the window lower and compact while the session is still sharp.