Please address the comments from this code review:

## Overall Comments
- In `xtask/src/lib.rs`, `Exit::from(Cmd)` only surfaces `stderr` on non-zero exit and discards `stdout`; consider including `stdout` (or at least a truncated version) in the error message to make diagnosing failing commands easier.
- The xtask `main` currently anchors all commands at `Path::new(".")`; if this is intended to operate at the workspace root, consider resolving the workspace root explicitly (e.g., via `CARGO_MANIFEST_DIR` or `cargo metadata`) to avoid surprising behavior when run from subdirectories.

## Individual Comments

### Comment 1
<location path="xtask/tests/test_fmt.rs" line_range="14-15" />
<code_context>
+    let original = fs::read_to_string("tests/fixture/src/lib.rs").unwrap();
+    let copied = fs::read_to_string(tmp.path().join("src/lib.rs")).unwrap();
+    assert_eq!(original, copied);
+    let _ = fmt(tmp.path());
+    let formatted = fs::read_to_string(tmp.path().join("src/lib.rs")).unwrap();
+    assert_ne!(original, formatted);
+}
</code_context>
<issue_to_address>
**suggestion (testing):** Assert that `fmt` actually succeeds instead of ignoring its result, so failures in running `cargo fmt` don’t show up as confusing assertion errors.

`fmt(tmp.path())` is currently ignored, so if `cargo fmt` fails (e.g. missing rustfmt, bad Cargo.toml), the test only fails later at `assert_ne!(original, formatted)`, which is misleading.

Instead, check the command result before reading the file, e.g.:

```rust
let cmd = fmt(tmp.path());
let output = cmd.result.expect("`cargo fmt` failed to run");
assert!(
    output.status.success(),
    "`cargo fmt` exited with status {:?}\nstderr: {}",
    output.status,
    String::from_utf8_lossy(&output.stderr),
);
```

That way the test fails clearly when `cargo fmt` itself fails, and only then checks that the file changed.
</issue_to_address>