fix(tracker): close alpha accounting review gaps

Address post-review blockers from the alpha hardening diff: charge non-stream completions that omit usage, include prompt tokens in max_charge_per_request enforcement, and wire the CLI to a persisted local registry contract by default.

Verification: .venv/bin/python -m compileall -q packages tests; focused billing/CLI checks passed; full pytest: 314 passed, 3 skipped, 1 env failure from existing meshnet-node on port 7000.

tests/test_billing_ledger.py

@@ -21,6 +21,7 @@ from meshnet_tracker.server import (
     _billable_non_stream_tokens,
     _billable_stream_tokens,
     _estimate_prompt_tokens,
+    _observed_non_stream_completion_tokens,
     _observed_stream_tokens,
 )
 
@@ -99,6 +100,14 @@ def test_non_stream_billable_tokens_cap_usage_by_request_bound():
     assert _billable_non_stream_tokens(payload, request) == 5
 
 
+def test_non_stream_billable_tokens_fallback_when_usage_missing():
+    payload = {"choices": [{"message": {"role": "assistant", "content": "ok now"}}]}
+    request = {"messages": [{"role": "user", "content": "hello tracker"}], "max_tokens": 10}
+
+    assert _observed_non_stream_completion_tokens(payload) == 2
+    assert _billable_non_stream_tokens(payload, request) == 4
+
+
 def test_stream_billable_tokens_allow_usage_to_lower_observed_count_only():
     observed_payload = {
         "choices": [{
@@ -557,7 +566,12 @@ def test_proxy_chat_rejects_request_above_spend_cap_before_routing():
             r.read()
 
         with pytest.raises(urllib.error.HTTPError) as exc_info:
-            _chat(tracker_url, api_key="capped-client", max_tokens=1000)
+            _chat(
+                tracker_url,
+                api_key="capped-client",
+                messages=[{"role": "user", "content": " ".join(["prompt"] * 1000)}],
+                max_tokens=1,
+            )
         body = json.loads(exc_info.value.read())
         assert exc_info.value.code == 402
         assert body["error"]["code"] == "spend_cap_exceeded"