Compact Your Context at 60%, Not 95%
Most people let their context window fill to 95% before doing anything about it. By then, Claude is already degrading -- repeating itself, losing earlier context, making avoidable mistakes.
A better approach: check early, act early.
/context # see exactly where your tokens are going
/compact # summarise when you hit around 60%
/clear # wipe completely between unrelated tasks
The 60% threshold is the key insight. At 60%, /compact still captures everything important because Claude has full visibility over the entire session. At 95%, it is doing triage -- and you will feel it in the output quality.
A few habits that compound the benefit:
- Never paste 200+ lines of unrelated code into a prompt -- it silently crowds out your actual task context
- Always
/clearwhen switching between unrelated tasks, even if you are well under the limit - Run
/contextbefore/compactto understand what is eating your tokens, not just how much
The individual commands (/context, /compact, /clear) are quick to learn. The discipline of using them proactively -- before things degrade, not after -- is what actually changes your results. Think of 60% as a maintenance interval, not a warning sign.
Treat context management like memory hygiene: small, regular actions beat emergency cleanup every time.
via @Hartdrawss
Log in to leave a comment.
A PreToolUse hook can intercept test runner commands and filter output to show only failures, cutting thousands of tokens from Claude's context.
CLAUDE.md loads into every message. Move workflow-specific instructions into skills that load on demand to reduce token costs across your session.
Every event emitted while processing a single prompt shares a prompt.id UUID, letting you trace the complete chain of API calls and tool executions.