PyPI page
Home page
Author:
Ondřej Úlehla
Summary:
Eval/replay harness for AI agents: run agent tasks in isolated E2B sandboxes, record full traces, measure pass rate, flakiness and cost.
Latest version:
0.1.0
Required dependencies:
e2b
|
pydantic
|
python-dotenv
|
pyyaml
|
rich
|
typer
Downloads last day:
0
Downloads last week:
3
Downloads last month:
17
7-day smoothing