Filter Test Output with a PreToolUse Hook to Cut Token Costs

Running pytest or npm test can dump thousands of lines into Claude's context. Most of that output is passing tests you don't care about. A PreToolUse hook can intercept test commands and pipe them through a filter before Claude ever sees the results.

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Bash",
        "hooks": [
          {
            "type": "command",
            "command": "~/.claude/hooks/filter-test-output.sh"
          }
        ]
      }
    ]
  }
}

The hook script checks if the command is a test runner and rewrites it to only show failures:

#!/bin/bash
input=$(cat)
cmd=$(echo "$input" | jq -r '.tool_input.command')

if [[ "$cmd" =~ ^(npm test|pytest|go test) ]]; then
  filtered_cmd="$cmd 2>&1 | grep -A 5 -E '(FAIL|ERROR|error:)' | head -100"
  echo "{\"hookSpecificOutput\":{\"hookEventName\":\"PreToolUse\",\"permissionDecision\":\"allow\",\"updatedInput\":{\"command\":\"$filtered_cmd\"}}}"
else
  echo "{}"
fi

This can reduce context from tens of thousands of tokens to a few hundred. Claude still sees all the failures it needs to debug, just none of the noise.

Less noise in, fewer tokens burned, same debugging power.

via Claude Code Docs — Costs