ci: add the two repaired tests to the Windows lane

apps/desktop/package.json

@@ -52,7 +52,7 @@
     "fix": "npm run lint:fix && npm run fmt",
     "test:ui": "vitest run --project ui",
     "test:desktop:platforms": "vitest run --project electron",
-    "test:desktop:win-install": "vitest run --project electron desktop-electron-pin assert-dist-built assert-win-vcruntime write-build-stamp",
+    "test:desktop:win-install": "vitest run --project electron desktop-electron-pin assert-dist-built assert-win-vcruntime write-build-stamp update-relaunch windows-hermes-path",
     "test": "vitest run",
     "preview": "node scripts/assert-root-install.mjs && vite preview --host 127.0.0.1 --port 4174",
     "check:test:desktop:platforms": "npm run test:desktop:platforms",

docs/system-log/2026-08-09.md

@@ -331,3 +331,30 @@ I should have checked the diff before naming a culprit.
 - **Follow-up:** the two remaining items are unchanged and both are somebody
   else's call — `electron-builder` is unimportable under vitest on Windows
   (likely upstream), and Windows' no-mux SSH path still has no coverage.
+
+## 2026-08-09T19:10:00Z — Add the two repaired tests to the Windows lane (Claude Code)
+
+- **Agent/tool:** Claude Code
+- **Repos:** hermes-agent
+- **Done:**
+  - Held back from the 18:40 change on purpose; now that both files are
+    cross-platform and merged, added `update-relaunch` and
+    `windows-hermes-path` to `test:desktop:win-install`. The lane goes from
+    four files to six.
+  - **Evidence that the rest of both files is Windows-safe, which is the thing
+    I got wrong twice before.** These are whole files — 16 and 22 tests — not
+    just the two assertions repaired at 18:40. In the first full-suite Windows
+    run (job 93652696640) the only failures from these two files were
+    `update-relaunch.test.ts:54` and `windows-hermes-path.test.ts:177`; every
+    other test in them was **observed passing on a Windows runner**. The
+    remaining ssh-connection failures are in a file that stays excluded.
+  - **The one gap observation could not cover:** vitest stops at the first
+    failing assertion in a test, so the second assertion of the test at :53
+    was never reached on Windows. The win32 simulation covers it explicitly
+    ("unpacked itself") and it passes after the repair — 14 passed, 0 failed
+    across posix and win32.
+  - Re-verified the filter selection: six of the project's 80 test files, with
+    `before-pack`, `stage-native-deps` and `ssh-connection` still excluded.
+- **Follow-up:** unchanged, and both are outside my call — `electron-builder`
+  is unimportable under vitest on Windows (likely upstream), and Windows'
+  no-mux SSH path still has no coverage.