fix(test): fix sync continuation test mock leaking across sessions

The messages() mock in 'session_id with background=false' test did not
filter by session ID, causing resolveParentContext's SDK calls for
parent-session to increment messagesCallCount. This inflated
anchorMessageCount to 4 (matching total messages), so the poll loop
could never detect new messages and always hit MAX_POLL_TIME_MS.

Fix: filter messages() mock by path.id so only target session
(ses_continue_test) increments the counter. Restore MAX_POLL_TIME_MS
from 8000 back to 2000.

src/tools/delegate-task/tools.test.ts

@@ -45,7 +45,7 @@ describe("sisyphus-task", () => {
       STABILITY_POLLS_REQUIRED: 1,
       WAIT_FOR_SESSION_INTERVAL_MS: 10,
       WAIT_FOR_SESSION_TIMEOUT_MS: 1000,
-      MAX_POLL_TIME_MS: 8000,
+      MAX_POLL_TIME_MS: 2000,
       SESSION_CONTINUATION_STABILITY_MS: 50,
     })
     cacheSpy = spyOn(connectedProvidersCache, "readConnectedProvidersCache").mockReturnValue(["anthropic", "google", "openai"])
@@ -1273,7 +1273,13 @@ describe("sisyphus-task", () => {
          session: {
            prompt: async () => ({ data: {} }),
            promptAsync: async () => ({ data: {} }),
-          messages: async () => {
+           messages: async (args?: { path?: { id?: string } }) => {
+             const sessionID = args?.path?.id
+             // Only track calls for the target session (ses_continue_test),
+             // not for parent-session calls from resolveParentContext
+             if (sessionID !== "ses_continue_test") {
+               return { data: [] }
+             }
              messagesCallCount++
              const now = Date.now()