JevHarness
LLM-authored task-specific Jev harnesses with optional full-trajectory reward reflection and GEPA evolution.
How it uses Jev
Let an LLM write a task-specific harness for Jev. Run it, inspect its decisions, and optionally improve it using rewards and complete execution traces.
- Languages
- Python 86% · HTML 10% · JavaScript 4%
- Last activity
- Sep 2026
- Added
- 2026-09-21
- Stars at snapshot
- 2
- Status
- Community listing
Source: community catalog. Details and metrics reflect the published community snapshot and are not a performance endorsement. Documentation details extracted from GitHub · the README · the project website.