fix: address independent verification findings

Classify empty completions as transient per human ruling (fixes flaky
real-gateway smoke and prevents caching empty responses), rename the
factory injection parameter gate to breaker per the frozen design,
rewrite the probe-entry cleanup without except BaseException, declare
python-dotenv explicitly, add a mid-backoff cancellation test, and
record all implementation errata in the design and architecture docs.

src/polygateway/transports/openai_compat.py

@@ -252,6 +252,7 @@ class OpenAICompatTransport:
                 (content_parts if is_content else thinking_parts).append(text)
         salvaged = self._check_done(sink, content_parts, thinking_parts, source)
         content, thinking = self._finalize_text(content_parts, thinking_parts, profile)
+        self._reject_empty_completion(content, source)
         prompt, completion, usage_source = _resolve_usage(sink.get("usage") or {}, source)
         if salvaged:
             usage_source = "estimated"  # 打捞路径强制 estimated(设计 §6)
@@ -283,6 +284,20 @@ class OpenAICompatTransport:
             raise TransientError(f"{source.name} SSE missing_done: 截断且无 [DONE]", **ctx)
         return True
 
+    def _reject_empty_completion(self, content: str, source: SourceConfig) -> None:
+        """空补全 → 瞬时错误(2026-07-20 人类裁决,M1 验证发现)。
+
+        服务 200 且流程完整([DONE]/usage 正常)但 content 为空——MiniMax 等
+        网关的间歇异常形态。视为服务抖动: 退避重试/换源,**绝不缓存空响应**;
+        承 CHS "VLM 零 content"归瞬时的先例(invokers.py:309)。
+        """
+        if not content.strip():
+            raise TransientError(
+                f"{source.name} 空补全(empty_completion): 流程完整但零内容",
+                source_name=source.name,
+                operation="chat",
+            )
+
     def _finalize_text(
         self, content_parts: list[str], thinking_parts: list[str], profile: ProviderProfile
     ) -> tuple[str, str]:
@@ -321,6 +336,7 @@ class OpenAICompatTransport:
         content, thinking = self._finalize_text(
             [message.get("content") or ""], [message.get("reasoning_content") or ""], profile
         )
+        self._reject_empty_completion(content, source)
         prompt, completion, usage_source = _resolve_usage(body.get("usage") or {}, source)
         return TransportResult(
             content=content,