Hi, I'd like to suggest LongHorizon-Harness for inclusion in awesome-claude-skills.
Repo:
https://github.com/AMAP-ML/LongHorizon-Harness
Why it seems relevant here:
- open-source Python harness for long-running GUI and CLI tasks
- runs around Claude Code or Codex instead of replacing their native agent loops
- separates Manager, fresh-context Executor, and independent Auditor roles
- keeps only verified progress in durable task state
Suggested category: Coding agents
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
If this fits the list, I can prepare a clean README-only contribution following the repository format.
Hi, I'd like to suggest LongHorizon-Harness for inclusion in
awesome-claude-skills.Repo:
https://github.com/AMAP-ML/LongHorizon-Harness
Why it seems relevant here:
Suggested category:
Coding agentsI 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
If this fits the list, I can prepare a clean README-only contribution following the repository format.