#668 · Primary category: AI Agents & Automation

LongHorizon-Harness

agent claude claude-code claude-plugin cli codex codex-desktop codex-plugin cua deepseek-harness gui harness long-horizon long-horizon-agents longhorizon-harness loop loop-engineering

LongHorizon-Harness enables AI agents to run for dozens of hours across desktop apps and the CLI with preserved state, verified progress, and automatic recovery.

Project last updated:08/20/26

GitHub Stars

1.4K

Forks

142

Contributors

10

License

MIT

Why we included this project

Most long-running computer-use tasks die from context bloat or a single bad action, but LongHorizon-Harness tackles that by reshaping the execution loop around the coding agents you already use, whether that is Claude Code, Codex, OpenCode, or DeepSeek Harness. It keeps task state durable and verified, checks each step against the real desktop or terminal, and checkpoints accepted progress so you can recover from a failure and resume with fresh context instead of losing everything. That is what teams automating long workflows across GUI apps and the CLI need: progress that survives crashes, restarts, and context limits rather than a single-shot automation run.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category