feat: release v1.0.1 CADWorld 网站与 LeKiwi 智能抓放

集成同源 BYOK 会话隔离、精简模型设置、官方订阅入口和 HTTPS 发布运维；保留本地训练/调参与控制能力。同步 npm 版本及 CHANGELOG，记录公网真实 API 验收仍待用户凭据。

decision_server/providers/http.py

@@ -0,0 +1,63 @@
+"""Bounded, non-redirecting HTTP. No SDK retries or environment credential discovery."""
+
+import asyncio
+import math
+
+import aiohttp
+
+from ..protocol import DecisionError, loads
+
+
+async def post(session, connection, path, payload):
+    try:
+        async with session.post(
+            connection.base_url + path,
+            json=payload,
+            headers={"Authorization": "Bearer " + connection.key},
+            allow_redirects=False,
+            timeout=aiohttp.ClientTimeout(total=45),
+        ) as response:
+            if response.status != 200:
+                raise DecisionError(f"upstream_http_{response.status}", 502)
+            # read(n) may return a partial chunk: accumulate with an explicit byte bound.
+            data = bytearray()
+            async for chunk in response.content.iter_chunked(16384):
+                data.extend(chunk)
+                if len(data) > 262144:
+                    raise DecisionError("upstream_response_too_large", 502)
+            try:
+                result = loads(data.decode("utf-8"))
+            except UnicodeError as exc:
+                raise DecisionError("invalid_upstream_encoding", 502) from exc
+            if not isinstance(result, dict):
+                raise DecisionError("invalid_upstream_response", 502)
+            return result
+    except TimeoutError as exc:
+        raise DecisionError("upstream_timeout", 504) from exc
+    except (aiohttp.ClientError, OSError) as exc:
+        raise DecisionError("upstream_transport_error", 502) from exc
+    except asyncio.CancelledError:
+        raise
+
+
+def usage(result):
+    """Only real numeric counters, no inferred price, upstream strings or raw body."""
+    raw = result.get("usage", {})
+    if not isinstance(raw, dict):
+        return {}
+    return {
+        k: v
+        for k, v in raw.items()
+        if k
+        in (
+            "input_tokens",
+            "output_tokens",
+            "total_tokens",
+            "prompt_tokens",
+            "completion_tokens",
+            "cost",
+        )
+        and type(v) in (int, float)
+        and math.isfinite(v)
+        and 0 <= v <= 1e9
+    }