Introducing HarnessGrad
HarnessGrad is the gradient for harness self-improvement: the signal a method needs before it can improve anything, built so the number can be checked rather than asserted. This site is its front page.
What it is for
Section titled “What it is for”Every harness-evolution method reports a final score delta. None of them can say what that delta is made of, and none report it on a curve a reader can compare against another method’s. HarnessGrad defines one: a base harness, a fixed task set, a sampling loop, and a curve point that records the harness state, the two models involved, the score with an interval, and the compute spent.
The repository proposes no RSI method. It ships runnable reference methods, but the contribution is the instrument that decides whether any of them worked — and it measures every candidate itself, because a subject that scores itself makes the comparison meaningless.
What runs today
Section titled “What runs today”- The local demo, end to end, with zero network and zero cost.
- The measurement contract, written down as
INTERFACE.md. - Enforced isolation: a harness cannot see the platform while it runs, and runs made under different isolation rules are never placed on one axis.
- Per-dataset tests, so one change does not require the whole suite.
What is not claimed
Section titled “What is not claimed”Local results are not production evidence, and a demo or diagnostic grants no production authorization. Per-dataset gaps in runtime, quota, snapshot and cleanup enforcement are recorded with a disposition rather than smoothed over.
No dataset has been chosen, and the metric is defined no further than “defined on a task set and pairable”. Contamination — whether the agent model has seen the benchmark tasks — is not detectable from inside the framework, and is the largest known threat to the validity of any curve.
Getting started runs the demo. Status and limits is worth reading first.