Skip to content

Releases

v0.1.0 — Research preview

Pre-release

Choose a tag to compare

@imom39a imom39a released this 22 Sep 16:00

Research preview of the standalone Hanoi experiments: Jev alone, LLM alone, and LLM proposals checked by Jev with one repair attempt.

Includes two browser views, a bounded evaluator, an offline audit, six live pilot episodes with full traces, screenshots, and reproduction instructions. Actual target attainment is recorded separately from participant-accepted completion.

The pilot does not demonstrate a general hybrid advantage. A persistent LLM planner with Jev execution is documented as proposed research, not implemented behavior.

Validation: 30 tests, package installation, documentation links, and dataset checksums passed locally.