Newborn 
PY A platform for running and comparing harness self-improvement methods.
https://zjunlp.github.io/HarnessGrad-Hub/ ↗ HEALTH ? 88 / Healthy
Newborn PY // readme
What you can run
A run is one point in a harness × method × benchmark space: a harness solves tasks, a method proposes edits to that harness between rounds, and a benchmark scores each candidate. HarnessGrad supplies the three middle layers that every RSI paper re-implements badly — tasks and scoring, an isolated sandbox, and version control over the candidate harness — so an experiment is one command and a curve.
| axis | what ships | where |
|---|---|---|
| harnesses | 8: loop, codex, opencode, deepseek-harness, openclaw, miniagent, miniswe, oracle | base_harness/<name>/ |
| RSI methods | 19: dgm, rrsi, adas, ahe, ace, skillopt, evoprompt, evo_harness, opro, tthe, dream_rsi, sica_ci, gepa, darwinx, arbor, hyperagents, meta_harness, harnessx, modular_rsi | experiments/method-<name>.json |
| benchmarks | 56 declared, 56 runnable, none refused — every one listed: agentdojo, aime, alfworld, appworld, authored_example_fix_bug, bbeh, bfcl_v3, bird_minidev, browsecomp, chessbench_puzzles, cooperbench, costbench, costbench_dynamic, deepsearchqa, deepswe, ds1000,… |