From d477341e6af816ff4396042b084a0378ef437b4a Mon Sep 17 00:00:00 2001 From: Ogulcan Celik Date: Sat, 30 May 2026 23:54:41 +0300 Subject: refactor: split agent detection by agent Move screen detection into per-agent modules and keep visible-state arbitration centralized in terminal state. Tighten Codex prompt-scoped working detection and Claude hook/screen recovery for stale blocked and idle reports. Add a focused just recipe for single-test runs. refs #349 refs #352 --- justfile | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'justfile') diff --git a/justfile b/justfile index 658d6d7a..8922e2fe 100644 --- a/justfile +++ b/justfile @@ -5,6 +5,10 @@ test: cargo nextest run --locked --status-level fail --final-status-level fail --failure-output final --success-output never python3 -m unittest scripts.test_changelog scripts.test_vendor_libghostty_vt +# Run one nextest filter, e.g. `just test-one codex_stale_working` +test-one filter: + cargo nextest run --locked "{{filter}}" --status-level fail --final-status-level fail --failure-output final --success-output never + # Run fast local lint checks lint: cargo fmt --check -- cgit v1.2.3