Where to look first

Read command.session?.diagnostics before anything else

When a CLI fails to start, the protocol stream carries a result saying the run errored, but the reason — a bad flag, a missing login, a version mismatch — is on stderr and nowhere else. A host that discards it is left with “it failed” and no way to say why.

console.error(command.session?.diagnostics);
// e.g. "No conversation found with session ID: abc"

And inspect() when the question is what was actually going to run:

console.log(command.inspect());
// provider, full argv, the stdin payload, model, permissions, session liveness

Why these two come first

diagnostics and inspect() answer the two halves of almost every failure.

Question	What to read
Why did the CLI refuse or die?	command.session?.diagnostics — stderr, in full
What was going to run at all?	command.inspect() — provider, argv, stdin payload, model, permissions, session state

inspect() is useful even before anything has run, because nothing reaches the CLI until send(). A command that has been assembled but not sent still describes itself completely, ending with (not started).

const command = Command.open(Claude, { workingDir: '/proj' });
command.setModel('opus').setPermissions('plan').setMessage('Plan the migration');

console.log(command.inspect());
// provider:    claude
// command:     claude -p --output-format stream-json …
// stdin:       {"type":"user","message":{"role":"user","content":"Plan the migration"}}
// model:       opus
// permissions: plan
// attachments: 0
// workingDir:  /proj
// session:     3f2a… (not started)

toString() delegates to inspect(), so a command logged or interpolated reads as its full state rather than as [object Object].

Then the symptom index

Once you know what ran and what the CLI said about it, the symptom index maps what you are seeing to its most likely cause.