tangle-network/agent-eval

agent-eval

Maintain agent-eval cases, judges, records, traces, campaigns, comparisons, and releases.

Idioma do texto original: inglês

sweagent-traj

Parse classic SWE-agent .traj trajectories: one step per entry of the trajectory array, in order, matching the CodeTraceBench annotation convention. No upstream CodeTracer commit ships a SWE-agent parser.

Idioma do texto original: inglês

openhands-completions

Parse OpenHands trajectories published as per-call LiteLLM completion logs (<model>-<timestamp>.json files with messages/response/kwargs/cost), the layout of CodeTraceBench's swe_raw OpenHands trials. Reads the call file whose messages hold the most tool-call…

Idioma do texto original: inglês

openhands-sessions

Parse OpenHands session-event trajectories at the CodeTraceBench annotation granularity: one step per action event with a cause-paired observation (run, run_ipython, read, edit, think, recall). Outranks the seed openhands skill, which counts only…

Idioma do texto original: inglês

terminus2-commands

Parse Terminus2 trajectories at command granularity: one step per entry of each episode response's commands array, matching the CodeTraceBench annotation convention. Outranks the seed terminus2 skill, whose episode-level steps do not align with the…

Idioma do texto original: inglês