feat: Terminal-Bench 2.0 harness, Pi agent wrapper, and `vox bench` subcommand#2

Draft

Copilot wants to merge 3 commits into

mainfrom

copilot/add-terminal-bench-2-0-benchmark

Commits on Mar 13, 2026

Initial plan
Copilot
committed
feat: add Terminal-Bench 2.0 harness, Pi agent, and vox bench subcommand

Copilot
and
aryateja2106
committed
fix: address code review — move contextlib import to module level, fix docstring categories

Copilot
and
aryateja2106
committed