lazy-harness¶

A cross-platform harnessing framework for AI coding agents.

lazy-harness turns a raw AI coding agent (Claude Code today, others planned) into a daily-driver workstation by adding the scaffolding that agents do not ship with: multi-profile isolation, a hook engine, a monitoring pipeline, a knowledge directory, a scheduler, and a session memory model that persists across conversations.

What it gives you¶

Profiles. Isolate separate agent setups — personal, work, client, experimental — with their own CLAUDE.md, settings.json, skills, and knowledge. Switch by directory or env var.
Hooks. A cross-platform hook engine with built-in hooks for session-start context injection, pre-compact summaries, session export, and compound loop enforcement. Bring your own hooks via config.
Monitoring. SQLite-backed metrics on every session: duration, message count, tools used, cost. Nine built-in lh status views.
Knowledge. A filesystem knowledge directory for sessions and distilled learnings, optionally indexed by QMD for semantic search.
Scheduler. A unified interface over launchd, systemd, and cron. Declare recurring jobs in config.toml; lh scheduler install does the rest.
Migration. lh migrate takes any existing Claude Code setup and upgrades it into a lazy-harness installation with a dry-run gate and full rollback.

Quick start¶

uv tool install git+https://github.com/lazynet/lazy-harness
lh init                    # new install
# or
lh migrate --dry-run       # existing Claude Code setup
lh migrate
lh doctor                  # verify prerequisites
lh selftest                # verify the framework itself

See the getting-started guide for details.

Why this exists¶

Read the problem and the memory model.

Status¶

Framework v0.4.0 is the first stable release. Supported platforms: macOS, Linux. Supported agent: Claude Code (others planned via the adapter layer).