test(windows): stabilize shell suites and safe runner in WSL

tests/additional_edge_tests.sh

@@ -54,6 +54,9 @@ if [ -d "$PROJECT_DIR/src" ]; then
 fi
 TM="./tm"
 
+# Ensure orchestration enforcement allows non-interactive test commands.
+export TM_AGENT_ID="test_suite_agent"
+
 TEST_COUNT=0
 PASS_COUNT=0
 
@@ -65,7 +68,7 @@ cleanup() {
     # Add delay for WSL
     if [ $IS_WSL -eq 1 ]; then
         sleep 0.5
-        sync
+        sleep 0.1
     fi
 
     cd "$SAVED_DIR" 2>/dev/null || cd /tmp

tests/run_all_tests_safe.sh

@@ -112,7 +112,6 @@ run_test_safe() {
     if [ $IS_WSL -eq 1 ]; then
         echo "Waiting 2 seconds before next test (WSL safety)..."
         sleep 2
-        sync
     else
         sleep 0.5
     fi
@@ -163,4 +162,4 @@ if [ $FAILED_TESTS -eq 0 ] && [ $PASSED_TESTS -gt 0 ]; then
 else
     echo -e "${RED}Some tests failed. Please review the results.${NC}"
     exit 1
-fi
+fi

tests/stress_test.sh

@@ -54,6 +54,9 @@ if [ -d "$PROJECT_DIR/src" ]; then
 fi
 TM="./tm"
 
+# Ensure orchestration enforcement allows non-interactive stress commands.
+export TM_AGENT_ID="test_suite_agent"
+
 TEST_COUNT=0
 PASS_COUNT=0
 FAIL_COUNT=0
@@ -67,7 +70,7 @@ cleanup() {
     # Add delay for WSL
     if [ $IS_WSL -eq 1 ]; then
         sleep 0.5
-        sync
+        sleep 0.1
     fi
 
     cd "$SAVED_DIR" 2>/dev/null || cd /tmp
@@ -88,10 +91,16 @@ function stress_test() {
     local END_TIME=$(date +%s.%N)
     local DURATION=$(echo "$END_TIME - $START_TIME" | bc -l)
 
-    # Check if completed within reasonable time (adjust as needed)
-    if (( $(echo "$DURATION < 10.0" | bc -l) )); then
+    # WSL is slower under filesystem-heavy loops; use a higher threshold there.
+    local DURATION_LIMIT="10.0"
+    if [ $IS_WSL -eq 1 ]; then
+        DURATION_LIMIT="30.0"
+    fi
+
+    # Check if completed within reasonable time.
+    if (( $(echo "$DURATION < $DURATION_LIMIT" | bc -l) )); then
         if [ -n "$EXPECTED_COUNT" ]; then
-            local ACTUAL_COUNT=$($TM list | grep -c "○\|◐\|●\|⊘\|✗")
+            local ACTUAL_COUNT=$($TM export --format json 2>/dev/null | grep -o '"id"' | wc -l | tr -d ' ')
             if [ $ACTUAL_COUNT -ge $EXPECTED_COUNT ]; then
                 echo "PASS (${DURATION}s, $ACTUAL_COUNT tasks)"
                 PASS_COUNT=$((PASS_COUNT + 1))
@@ -303,7 +312,7 @@ echo "╔═══════════════════════
 echo "║                  STRESS TEST SUMMARY                 ║"
 echo "╚══════════════════════════════════════════════════════╝"
 
-TOTAL_TASKS=$($TM list | grep -c "○\|◐\|●\|⊘\|✗" 2>/dev/null || echo 0)
+TOTAL_TASKS=$($TM export --format json 2>/dev/null | grep -o '"id"' | wc -l | tr -d ' ' || echo 0)
 echo "Total tasks created: $TOTAL_TASKS"
 echo "Total stress tests: $TEST_COUNT"
 echo "Passed: $PASS_COUNT"
@@ -322,4 +331,4 @@ else
 fi
 
 # Cleanup is handled by trap
-exit $EXIT_CODE
+exit $EXIT_CODE

tests/test_agent_specialization.sh

@@ -42,6 +42,7 @@ if [ -d "$PROJECT_DIR/src" ]; then
 fi
 chmod +x ./tm
 TM="./tm"
+export TM_AGENT_ID="test_suite_agent"
 
 echo -e "${BLUE}===================================${NC}"
 echo -e "${BLUE}  Agent Specialization Tests       ${NC}"

tests/test_collaboration.sh

@@ -39,6 +39,7 @@ else
 fi
 
 chmod +x tm
+export TM_AGENT_ID="test_suite_agent"
 
 echo -e "${BLUE}===================================${NC}"
 echo -e "${BLUE}   Collaboration Features Tests    ${NC}"

tests/test_context_sharing.sh

@@ -61,11 +61,12 @@ setup() {
     fi
     TM="./tm"
     chmod +x "$TM"
+    export TM_AGENT_ID="test_suite_agent"
 
-    # Add WSL-specific delay for filesystem sync
+    # Add WSL-specific delay without host-wide sync (can hang on mounted filesystems)
     if [ $IS_WSL -eq 1 ]; then
         sleep 0.5
-        sync
+        sleep 0.1
     fi
 
     git init > /dev/null 2>&1
@@ -103,7 +104,7 @@ cleanup() {
     # Add delay for WSL
     if [ $IS_WSL -eq 1 ]; then
         sleep 0.5
-        sync
+        sleep 0.1
     fi
 
     cd / 2>/dev/null || cd /tmp

tests/test_durability.sh

@@ -42,6 +42,7 @@ if [ -d "$PROJECT_DIR/src" ]; then
 fi
 chmod +x ./tm
 TM="./tm"
+export TM_AGENT_ID="test_suite_agent"
 
 echo -e "${BLUE}===================================${NC}"
 echo -e "${BLUE}    Durability & Resilience Tests  ${NC}"
@@ -83,8 +84,8 @@ test_basic_persistence() {
     # Backup database
     cp .task-orchestrator/tasks.db .task-orchestrator/tasks.db.backup
 
-    # Simulate restart by clearing any cache (if exists)
-    sync
+    # Simulate restart without calling host-wide sync (can hang on mounted filesystems).
+    sleep 0.1
 
     # Verify data persists
     $TM show $TASK1 | grep -q "in_progress" && \
@@ -191,8 +192,8 @@ test_wal_recovery() {
     # Create tasks
     TASK=$($TM add "WAL test task" | grep -o '[a-f0-9]\{8\}')
 
-    # Force checkpoint
-    sync
+    # Force checkpoint simulation without host-wide sync.
+    sleep 0.1
 
     # Verify WAL files exist (if WAL enabled)
     if [ -f .task-orchestrator/tasks.db-wal ]; then

tests/test_edge_cases.sh

@@ -53,6 +53,9 @@ if [ -d "$PROJECT_DIR/src" ]; then
 fi
 TM="./tm"
 
+# Ensure orchestration enforcement allows non-interactive test commands.
+export TM_AGENT_ID="test_suite_agent"
+
 TEST_COUNT=0
 PASS_COUNT=0
 FAIL_COUNT=0
@@ -67,7 +70,7 @@ cleanup() {
     # Add delay for WSL
     if [ $IS_WSL -eq 1 ]; then
         sleep 0.5
-        sync
+        sleep 0.1
     fi
 
     cd "$SAVED_DIR" 2>/dev/null || cd /tmp

tests/test_tm.sh

@@ -27,10 +27,10 @@ fi
 echo "Running tests in isolated directory: $TEST_DIR"
 cd "$TEST_DIR"
 
-# Add WSL-specific delay for filesystem sync
+# Add WSL-specific delay without host-wide sync (can hang on mounted filesystems)
 if [ $IS_WSL -eq 1 ]; then
     sleep 0.5
-    sync
+    sleep 0.1
 fi
 
 # Find tm executable safely
@@ -54,6 +54,9 @@ if [ -d "$PROJECT_DIR/src" ]; then
 fi
 TM="./tm"
 
+# Ensure orchestration enforcement allows non-interactive test commands.
+export TM_AGENT_ID="test_suite_agent"
+
 # Set environment variables for WSL safety
 if [ $IS_WSL -eq 1 ]; then
     export SQLITE_TMPDIR="$TEST_DIR"
@@ -71,10 +74,10 @@ cleanup() {
     # Kill any remaining tm processes
     pkill -f "tm.*$TEST_DIR" 2>/dev/null || true
 
-    # Add delay for WSL
+    # Add delay for WSL without host-wide sync
     if [ $IS_WSL -eq 1 ]; then
         sleep 0.5
-        sync
+        sleep 0.1
     fi
 
     cd "$SAVED_DIR" 2>/dev/null || cd /tmp
@@ -319,4 +322,4 @@ else
 fi
 
 # Cleanup is handled by trap
-exit $EXIT_CODE
+exit $EXIT_CODE