context-break-evaluator

Installation
SKILL.md

Context Break Evaluator

Purpose

This skill teaches you how to assess whether continuing a conversation will produce better or worse results than starting a new one. Context management is a critical but often overlooked factor in getting high-quality responses from AI agents.

Core Concept

Every message added to a conversation increases context length. Longer context is not inherently better. Context becomes harmful when:

  • Prior tasks are complete and their details are no longer relevant
  • Failed attempts, corrections, and dead ends have accumulated
  • The topic has shifted substantially from where the conversation began
  • The model must sift through noise to find the signal

Starting a fresh conversation clears this noise. A well-timed context break often produces better results than continuing in a degraded session.

Evaluation Framework

Installs
1
GitHub Stars
2
First Seen
Sep 17, 2026