{"id":"harbor","name":"Harbor","maker":"Harbor Framework","url":"https://harborframework.com","repo":"harbor-framework/harbor","summary":"A framework for running and improving agents against evaluation tasks in containers, used by Terminal-Bench.","details":"Runs agents in sandboxed environments against benchmark tasks, collects results, and can generate trajectories for training. It supports running many tasks in parallel on cloud providers and works with several benchmarks. The code is open source under Apache 2.0.","open_source":true,"hosted":false,"also":["sandboxes"],"verified":"2026-11-15","category":"benchmarks","category_name":"Agent benchmarks","page":"https://modelmoot.com/tools/harbor","reviewed":"2026-11-15","checked":{"at":"2026-11-15T14:03:44.997Z","address_answers":true,"repository":{"at":"2026-11-15T23:25:51.807Z","archived":false,"licence":"Apache-2.0","language":"Python","last_change":"2026-11-14T06:38:08.000Z","latest_release":{"tag":"v0.26.0","published":"2026-11-13T23:49:24.000Z"}}},"note":"Descriptions are the site's own. Facts under \"checked\" were read automatically from the tool's address and public repository."}