Please address the comments from this code review:

## Overall Comments
- In `readOutput`, consider explicitly specifying a charset for `InputStreamReader` (e.g., `StandardCharsets.UTF_8`) instead of relying on the platform default to avoid surprises across different environments.
- The Windows/non-Windows command selection logic in the tests is duplicated across multiple test methods; consider extracting a small helper (e.g., `runEchoCommand` / `runCommand(String unixCmd, String windowsCmd)`) to keep the tests more concise and avoid repetition.

## Individual Comments

### Comment 1
<location path="capabilities-common/src/main/java/ai/wanaku/capabilities/sdk/common/ProcessRunner.java" line_range="45" />
<code_context>
-        String line;
-        while ((line = reader.readLine()) != null) {
-            output.append(line).append("\n");
+        try (BufferedReader reader = new BufferedReader(new InputStreamReader(process.getInputStream()))) {
+            StringBuilder output = new StringBuilder();
+            String line;
</code_context>
<issue_to_address>
**suggestion (bug_risk):** Consider specifying an explicit charset when reading process output.

`InputStreamReader` without a charset uses the platform default, which may differ from the subprocess encoding and vary between environments. Please pass an explicit charset (e.g., `new InputStreamReader(process.getInputStream(), StandardCharsets.UTF_8)`) to ensure consistent decoding across platforms.

Suggested implementation:

```java
import java.io.InputStreamReader;
import java.nio.charset.StandardCharsets;

```

```java
        try (BufferedReader reader = new BufferedReader(new InputStreamReader(process.getInputStream(), StandardCharsets.UTF_8))) {

```
</issue_to_address>

### Comment 2
<location path="capabilities-common/src/test/java/ai/wanaku/capabilities/sdk/common/ProcessRunnerTest.java" line_range="11-13" />
<code_context>

 public class ProcessRunnerTest {

+    private static boolean isWindows() {
+        return System.getProperty("os.name").toLowerCase().contains("win");
+    }
+
     @Test
</code_context>
<issue_to_address>
**suggestion:** Consider extracting OS-specific command construction to a helper to avoid duplication across tests

`runWithOutput_returnsCommandOutput`, `runWithOutput_nonZeroExitDoesNotThrow`, and `runWithOutput_capturesStderrViaMergedStream` all duplicate the `if (isWindows()) { cmd.exe ... } else { sh ... }` logic. A small helper like `private static String[] shellCommand(String script)` that returns the OS‑specific command array would centralize this logic so future changes to OS detection or shell invocation only need to be made in one place.

Suggested implementation:

```java
    private static boolean isWindows() {
        return System.getProperty("os.name").toLowerCase().contains("win");
    }

    private static String[] shellCommand(String script) {
        if (isWindows()) {
            return new String[]{"cmd.exe", "/c", script};
        } else {
            return new String[]{"sh", "-c", script};
        }
    }

```

```java
    @Test
    void runWithOutput_returnsCommandOutput() {
        String[] command = shellCommand("echo hello");
        String output = ProcessRunner.runWithOutput(command);

        assertNotNull(output, "Output should not be null");

```

There are likely similar OS-specific command constructions in the tests `runWithOutput_nonZeroExitDoesNotThrow` and `runWithOutput_capturesStderrViaMergedStream`. For each of those tests, replace the duplicated `if (isWindows()) { cmd.exe ... } else { sh ... }` blocks by constructing the command using `shellCommand("<appropriate script>")` and then calling `ProcessRunner.runWithOutput(command)` (or the relevant ProcessRunner method) with the resulting array.
</issue_to_address>