Hi, I'd like to suggest a complementary direction for open-codex-computer-use.
Project:
https://github.com/AMAP-ML/LongHorizon-Harness
Your project focuses on GUI and desktop interaction for agents.
LongHorizon-Harness focuses on carrying one task across many rounds without state drift. It keeps the goal and verified progress in durable state, starts each execution round with a fresh context, and uses an independent Auditor before accepting progress.
Why this may be useful here:
- A short workflow note could show how to avoid context drift and premature completion in a task that spans many rounds.
- it adds an outer task-lifecycle layer without claiming to replace the project's existing agent or runtime
If this direction fits, I can prepare a small documentation or example contribution following the repository's conventions.
I am part of the team behind LongHorizon-Harness, so I wanted to disclose that connection.
Paper:
https://arxiv.org/abs/2608.01964
Project page:
https://lh-harness.pages.dev
Hi, I'd like to suggest a complementary direction for
open-codex-computer-use.Project:
https://github.com/AMAP-ML/LongHorizon-Harness
Your project focuses on GUI and desktop interaction for agents.
LongHorizon-Harness focuses on carrying one task across many rounds without state drift. It keeps the goal and verified progress in durable state, starts each execution round with a fresh context, and uses an independent Auditor before accepting progress.
Why this may be useful here:
If this direction fits, I can prepare a small documentation or example contribution following the repository's conventions.
I am part of the team behind LongHorizon-Harness, so I wanted to disclose that connection.
Paper:
https://arxiv.org/abs/2608.01964
Project page:
https://lh-harness.pages.dev