Live·python · memory

long-task-harness

Preserve context across long projects spanning multiple coding agent sessions.


Overview

When working on complex projects with AI coding agents, context gets lost between sessions. This harness maintains continuity by tracking:

  • What's been done
  • What's left to do
  • Key decisions made
  • Blockers encountered

How It Works

The harness creates a structured task file that persists across sessions. Each time you start a new session, the agent reads this file to understand the current state and picks up where it left off.

Usage

Best used as a skill – the agent automatically maintains the task file as it works.