fix: tolerate unavailable CPU sensor telemetry

audit/internal/platform/sat_host_packs.go

@@ -139,13 +139,34 @@ echo "no new thermal throttling detected during this run"
 `
 }
 
+// cpuSensorsProbeScript preserves lm-sensors output in the SAT evidence while
+// distinguishing an unsupported local hwmon interface from a broken command.
+//
+// Some otherwise healthy servers expose thermal telemetry only through BMC/IPMI
+// (or need a platform-specific hwmon module that the current kernel does not
+// provide). lm-sensors returns 1 and prints "No sensors found!" in that case.
+// That must not turn a successful CPU stress test into a false failure. Other
+// non-zero exits remain failures: they can indicate a missing/broken sensors
+// binary or a real runtime problem.
+func cpuSensorsProbeScript() string {
+	return `output=$(sensors 2>&1)
+rc=$?
+printf '%s\n' "$output"
+if [ "$rc" -ne 0 ] && printf '%s\n' "$output" | grep -Fq 'No sensors found!'; then
+  echo 'CPU temperature telemetry is unavailable through lm-sensors; continuing without local hwmon readings.'
+  exit 0
+fi
+exit "$rc"
+`
+}
+
 func cpuSATJobs(durationSec int) []satJob {
 	return []satJob{
 		{name: "01-lscpu.log", cmd: []string{"lscpu"}},
-		{name: "02-sensors-before.log", cmd: []string{"sensors"}},
+		{name: "02-sensors-before.log", cmd: []string{"sh", "-c", cpuSensorsProbeScript()}},
 		{name: "02-thermal-throttle-before.log", cmd: []string{"sh", "-c", cpuThrottleBeforeScript()}, informational: true},
 		{name: "03-stress-ng.log", cmd: []string{"stress-ng", "--cpu", "0", "--cpu-method", "all", "--timeout", fmt.Sprintf("%d", durationSec)}, syncBracket: true},
-		{name: "04-sensors-after.log", cmd: []string{"sensors"}},
+		{name: "04-sensors-after.log", cmd: []string{"sh", "-c", cpuSensorsProbeScript()}},
 		{name: "05-thermal-throttle-check.log", cmd: []string{"sh", "-c", cpuThrottleCheckScript()}},
 	}
 }

audit/internal/platform/sat_test.go

@@ -6,6 +6,7 @@ import (
 	"os"
 	"os/exec"
 	"path/filepath"
+	"strconv"
 	"strings"
 	"testing"
 	"time"
@@ -719,6 +720,42 @@ func TestCPUSATJobsIncludeThrottleCheck(t *testing.T) {
 	if jobs[5].name != "05-thermal-throttle-check.log" || jobs[5].informational {
 		t.Fatalf("throttle-check job=%+v want non-informational so it can fail overall_status", jobs[5])
 	}
+	for _, index := range []int{1, 4} {
+		if jobs[index].cmd[0] != "sh" || jobs[index].cmd[1] != "-c" || !strings.Contains(jobs[index].cmd[2], "No sensors found!") {
+			t.Fatalf("sensors job=%+v want probe that tolerates unavailable local hwmon", jobs[index])
+		}
+	}
+}
+
+func TestCPUSensorsProbeScript(t *testing.T) {
+	run := func(t *testing.T, output string, rc int) ([]byte, error) {
+		t.Helper()
+		binDir := t.TempDir()
+		mockSensors := filepath.Join(binDir, "sensors")
+		script := "#!/bin/sh\nprintf '%s\\n' '" + output + "'\nexit " + strconv.Itoa(rc) + "\n"
+		if err := os.WriteFile(mockSensors, []byte(script), 0755); err != nil {
+			t.Fatalf("write mock sensors: %v", err)
+		}
+		t.Setenv("PATH", binDir+":"+os.Getenv("PATH"))
+		return exec.Command("sh", "-c", cpuSensorsProbeScript()).CombinedOutput()
+	}
+
+	t.Run("no hardware-monitoring device is non-fatal", func(t *testing.T) {
+		out, err := run(t, "No sensors found!", 1)
+		if err != nil {
+			t.Fatalf("probe failed: %v; output=%s", err, out)
+		}
+		if !strings.Contains(string(out), "telemetry is unavailable") {
+			t.Fatalf("output=%s want unavailable telemetry annotation", out)
+		}
+	})
+
+	t.Run("unexpected sensors failure remains fatal", func(t *testing.T) {
+		out, err := run(t, "i2c read failed", 1)
+		if err == nil {
+			t.Fatalf("probe unexpectedly succeeded; output=%s", out)
+		}
+	})
 }
 
 // TestCPUThrottleCheckDetectsIncreaseDuringRun runs the real before/check