feat(cli): `coder-eval execute` + detached grading via `evaluate <run_dir>` - #154

Merged

akshaylive merged 17 commits into

mainfrom

akshaya/coder_eval_execute

Sep 9, 2026

Commits on Sep 3, 2026

feat(cli): add `coder-eval execute` — run tasks without grading them

akshaylive
and
claude
committed
0ee1b7d
feat(cli): grade an executed run afterwards — `evaluate <run_dir>` + `Sandbox.adopt`

akshaylive
and
claude
committed
39db361
feat(cli): make --resume distinguish "executed" from "graded"

akshaylive
and
claude
committed
7c13418
fix(eval): close the verdict-correctness gaps in detached grading

akshaylive
and
claude
committed
e831e92
fix(eval): address the medium and low findings from the branch review

akshaylive
and
claude
committed
489383d
test: fix two CI-only failures in the detached-grading tests

akshaylive
and
claude
committed
e9dfb94
Merge remote-tracking branch 'origin/main' into akshaya/coder_eval_execute
akshaylive
committed
698e213

Commits on Sep 4, 2026

fix(execute): close the verdict-changing and trust-boundary defects in detached grading

akshaylive
and
claude
committed
7d5d55d
fix(container): arm the heartbeat watchdog only inside the container

akshaylive
and
claude
committed
4f38a9e
fix(execute): close the verdict-divergence, trust-gate and fabricated-rate defects

akshaylive
and
claude
committed
19c60e4
fix(security): close the three CodeQL findings on the detached-grading diff

akshaylive
and
claude
committed
1652f40
fix(tests): kill the CodeQL taint source and the Windows mode assertion

akshaylive
and
claude
committed
9e479ae

Commits on Sep 9, 2026