[yaacovcorcos]

What Changed

packages/ssh/src/tunnel.test.ts, "replaces a managed server that exited but was not reaped": instead of reading the background child's process state once, the test now waits (polling every 0.1 s, at most ~10 s) until

1. the parent shell has become sleep, and only then kills the child, and
2. ps reports the child as a zombie (Z).

Test-only, inside a SCIENT-FORK region (the test comes from upstream T3 #19432; upstream main still has the same code). No timeouts raised, product code unchanged.

Why

The test is a recurring flake in the required "Test Workspaces" job: expected '' to include 'Z' (main runs 39170121279 and 39155611850, 18 October). A flake there throws PRs out of the merge queue.

The parent is sh -c "sleep 600 & echo $!; exec sleep 600". On Ubuntu sh is dash, which reaps exited children after every builtin, including that echo (dash 0.5.12 trap.c:90-91 installs the SIGCHLD handler; eval.c:922 reaps after builtins). The PID reaches the test while dash is still finishing the echo; a kill in that window is reaped before the exec, so no zombie exists and ps prints nothing, which is exactly the '' in the failing runs (a merely late kill would show a live state such as S).

Verification

• vp fmt --check, vp lint, tsc-rs --noEmit for packages/ssh, git diff --check: clean.
• On Windows the file runs 26 passed / 12 skipped; this test skips itself there, so this PR's Linux CI is its first real run. The polling loop's success and give-up exit codes were checked in Git Bash.

Documentation Impact

None — test-only change.

🤖 Generated with Claude Code

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

ℹ️ The exact PR base did not have a successful artifact. Baseline uses the latest successful main measurement shown below.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	5.6 KiB	5.6 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	4.3 KiB	4.3 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.3 KiB	1.3 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	26.2 KiB	26.2 KiB	0 B (0.0%)	29.3 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.6 KiB	5.6 KiB	0 B (0.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	4.3 KiB	4.3 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.3 KiB	1.3 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	26.6 KiB	26.6 KiB	0 B (0.0%)	29.3 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 2cce7fe · PR result: 94581a0 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.
Full compact HTTP snapshot with live cursor.

• Codex decoded thread snapshot: 133.1 KiB
• Claude decoded thread snapshot: 133.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.