coding-proxy 0.5.1a1__tar.gz → 0.5.1a3__tar.gz

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/AGENTS.md

@@ -39,15 +39,15 @@
   3. **Link Validity**: 确保所有引用的 URL 可访问且具备明确的上下文价值；
   4. **Testing**: 统一在 tests/ 下维护测试用例，区分单元测试（unit）和集成测试（integration），所有测试的本地运行总时间控制在 3 min 以内；
   5. **Pre-commit Hooks**: 首次克隆仓库使用 `uv run pre-commit install` 激活本地 Git hooks，使 Ruff lint（含 auto-fix）、Ruff format 及通用代码卫生检查在每次 commit 前自动运行。若 hooks 自动修复了问题，提交会被中断，执行 `git add -p` 审阅修复内容后重新提交即可；
-  6. **Issue**: 在 [issue.md](docs/agents/issue.md) 中维护你处理过的 Issue 摘要（问题描述、表因根因、处理方式、后续防范、同类问题影响与处理注意事项等），便于同类问题的跨上下文处理；注意识别相同 Issue，不要同 Issue 多处维护；
+  6. **Issue**: 在 [issue.md](docs/.agents/issue.md) 中维护你处理过的 Issue 摘要（问题描述、表因根因、处理方式、后续防范、同类问题影响与处理注意事项等），便于同类问题的跨上下文处理；注意识别相同 Issue，不要同 Issue 多处维护；
 - **Package Management Standardization (包管理规范)**:
   1. **Python**: 严禁使用 pip/poetry，**必须**统一使用 `uv` 进行包管理与脚本执行（如 `uv run`）；
   2. **JavaScript/TypeScript**: 严禁使用 npm/yarn，**必须**统一使用 `pnpm` 进行包管理与脚本执行；
 - **Database Management**: 谨慎操作，数据迁移、测试等操作严禁将现有数据删除，谨慎操作数据迁移的回滚，防止数据被清理。
 - **In-depth and close to the facts**：系统且全面地进行问题的分析，深入贴近事实，如有疑问，需先发问，不要乱做决定。
-- **Browser Validation Protocol (浏览器验证准则)**：Agent 不得自行完成、绕过或模拟任何 OAuth / SSO 认证流程，所有登录态均来源于用户已认证的 Chrome 主 profile（真实用户登录态）。完整协议（连通性自检、凭证管理、E2E 集成、实机回归等）详见 [浏览器验证协议](./docs/agents/browser-validation.md)；
+- **Browser Validation Protocol (浏览器验证准则)**：Agent 不得自行完成、绕过或模拟任何 OAuth / SSO 认证流程，所有登录态均来源于用户已认证的 Chrome 主 profile（真实用户登录态）。完整协议（连通性自检、凭证管理、E2E 集成、实机回归等）详见 [浏览器验证协议](./docs/.agents/browser-validation.md)；
   1. **安全红线**：禁止在 Sandbox 浏览器中跳转 Google 同意屏；禁止以模拟用户或第三方账号替代真实登录态；禁止要求用户在 chat 中粘贴密码、Cookie 或验证码；
-- **Knowledge Map (知识索引)**：项目所有文档索引统一维护在 [知识索引](./docs/agents/knowledge-map.md)，并在文档目录变更时即时同步跟新；
+- **Knowledge Map (知识索引)**：项目所有文档索引统一维护在 [知识索引](./docs/.agents/knowledge-map.md)，并在文档目录变更时即时同步跟新；
 - **Documentation Standards (文档规范)**：
   1. **Visual Documentation (图文并茂)**: 对于复杂逻辑，优先 **Mermaid Visualization Norms (Mermaid 可视化规范)**，构建”图文并茂”的直观文档；
      - **色彩语义与兼容性**：为图表节点配置具备语义辨识度的色彩，并确保在深色模式（Dark Mode）下具有极高的对比度与清晰度；
@@ -55,4 +55,4 @@
   2. **语言叙事**：用语精准，叙事完备，行文专业，聚焦核心，篇幅精炼，形象具体，体现真实作用与用户吸引性，字数恰当；
   3. **Direct Hyperlinking (直接跳转)**: 在文档中提及 Repo 内其他资源（文档/代码）时，**必须**构建可跳转的相对路径链接（如 `[Doc Name](./path.md)`），严禁使用”死文本”引用，以降低信息检索熵；
   4. **实操截图**：文档需要引入必要的浏览器实操截图时，需自行通过默认浏览器打开相关页面，通过实操现场截图并保留到文档路径进行文档引用；
-- **Reference Specifications (IEEE)**：为保障工程决策的可追溯性与学术严谨性，核心引用需遵循 [reference-specifications.md](docs/agents/reference-specifications.md)IEEE 标准引用格式；
+- **Reference Specifications (IEEE)**：为保障工程决策的可追溯性与学术严谨性，核心引用需遵循 [reference-specifications.md](docs/.agents/reference-specifications.md)IEEE 标准引用格式；

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: coding-proxy
-Version: 0.5.1a1
+Version: 0.5.1a3
 Summary: A High-Availability, Transparent, and Smart Multi-Vendor Proxy for Claude Code. Support Claude Plans, GitHub Copilot, Google Antigravity, ZAI/GLM, MiniMax, Qwen, Xiaomi, Kimi, Doubao...
 Project-URL: Source Code, https://github.com/ThreeFish-AI/coding-proxy
 Project-URL: User Guide, https://github.com/ThreeFish-AI/coding-proxy/blob/master/docs/user-guide.md

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "coding-proxy"
-version = "0.5.1a1"
+version = "0.5.1a3"
 description = "A High-Availability, Transparent, and Smart Multi-Vendor Proxy for Claude Code. Support Claude Plans, GitHub Copilot, Google Antigravity, ZAI/GLM, MiniMax, Qwen, Xiaomi, Kimi, Doubao..."
 readme = "README.md"
 requires-python = ">=3.12"

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/src/coding/proxy/config/config.default.yaml

@@ -339,6 +339,12 @@ model_mapping:
 # 未配置定价的模型在统计中 Cost 列显示 "-"
 pricing:
   # ── Anthropic Claude ──
+  - vendor: anthropic
+    model: claude-opus-4-8
+    input_cost_per_mtok: $5.0
+    output_cost_per_mtok: $25.0
+    cache_write_cost_per_mtok: $6.25
+    cache_read_cost_per_mtok: $0.50
   - vendor: anthropic
     model: claude-opus-4-7
     input_cost_per_mtok: $5.0

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/src/coding/proxy/convert/vendor_channels.py

@@ -52,6 +52,117 @@ def get_transition_channel(
 # ── 共享辅助函数 ──────────────────────────────────────────────
 
 
+def _dump_message_digest(
+    messages: list[dict[str, Any]],
+    *,
+    max_messages: int = 10,
+    label: str = "",
+) -> None:
+    """输出前 N 条消息的结构摘要（role + content_type_counts），用于过渡管线诊断.
+
+    仅在 DEBUG 级别输出，且仅在消息数 > 0 时才输出，避免噪声。
+    """
+    if not messages or not logger.isEnabledFor(logging.DEBUG):
+        return
+    parts: list[str] = [f"[{label}]" if label else ""]
+    limit = min(len(messages), max_messages)
+    for idx in range(limit):
+        msg = messages[idx]
+        role = msg.get("role", "?") if isinstance(msg, dict) else "?"
+        content = msg.get("content") if isinstance(msg, dict) else None
+        if isinstance(content, list):
+            type_counts: dict[str, int] = {}
+            for b in content:
+                if isinstance(b, dict):
+                    t = b.get("type", "?")
+                    type_counts[t] = type_counts.get(t, 0) + 1
+                else:
+                    type_counts["raw"] = type_counts.get("raw", 0) + 1
+            counts_str = ",".join(f"{t}:{c}" for t, c in sorted(type_counts.items()))
+        elif isinstance(content, str):
+            counts_str = f"str({len(content)})"
+        else:
+            counts_str = "empty"
+        parts.append(f"{idx}:{role}[{counts_str}]")
+    if len(messages) > max_messages:
+        parts.append(f"...+{len(messages) - max_messages}more")
+    logger.debug("Transition digest %s", " ".join(parts))
+
+
+def _validate_anthropic_pairing(
+    messages: list[dict[str, Any]],
+    *,
+    context: str = "",
+) -> list[str]:
+    """独立的 Anthropic tool_use/tool_result 配对自检（过渡管线末端执行）.
+
+    与 ``_enforce_pairing_sanity_pass`` 不同，此函数：
+    - 不修改消息列表（纯检测）
+    - 针对每个 assistant + tool_use，精确记录下一条 user 消息中匹配/缺失的 ID
+    - 发现不一致时输出 WARNING 级别日志含 message index 与具体 ID
+
+    Returns:
+        检测到的问题描述列表（空列表表示全部通过）。
+    """
+    issues: list[str] = []
+    for i, msg in enumerate(messages):
+        if not isinstance(msg, dict) or msg.get("role") != "assistant":
+            continue
+        content = msg.get("content")
+        if not isinstance(content, list):
+            continue
+        tool_use_ids = [
+            b["id"]
+            for b in content
+            if isinstance(b, dict) and b.get("type") == "tool_use" and b.get("id")
+        ]
+        if not tool_use_ids:
+            continue
+
+        next_idx = i + 1
+        if next_idx >= len(messages):
+            issues.append(f"messages[{i}]: assistant with tool_uses at end of list")
+            continue
+
+        next_msg = messages[next_idx]
+        if not isinstance(next_msg, dict) or next_msg.get("role") != "user":
+            issues.append(
+                f"messages[{i}]: next messages[{next_idx}] is not user "
+                f"(role={next_msg.get('role') if isinstance(next_msg, dict) else '?'})"
+            )
+            continue
+
+        user_content = next_msg.get("content")
+        if not isinstance(user_content, list):
+            user_content = []
+
+        result_ids = {
+            b["tool_use_id"]
+            for b in user_content
+            if isinstance(b, dict)
+            and b.get("type") == "tool_result"
+            and isinstance(b.get("tool_use_id"), str)
+        }
+
+        missing = [uid for uid in tool_use_ids if uid not in result_ids]
+        if missing:
+            issue = (
+                f"messages[{i}]: {len(missing)}/{len(tool_use_ids)} tool_use(s) "
+                f"without tool_result in messages[{next_idx}]: {missing[:5]}"
+            )
+            issues.append(issue)
+
+    if issues:
+        prefix = f"[{context}] " if context else ""
+        logger.warning(
+            "Anthropic pairing validation: %s%d issue(s) found: %s",
+            prefix,
+            len(issues),
+            "; ".join(issues),
+        )
+    return issues
+
+
 def strip_thinking_blocks(body: dict[str, Any]) -> int:
     """从 assistant 消息中移除 thinking/redacted_thinking 块（就地）.
 
@@ -678,6 +789,7 @@ def prepare_zhipu_to_anthropic(
     2. 改写 ``srvtoolu_*`` ID 与 ``server_tool_use`` 类型为标准 Anthropic 形式
     3. 强制 tool_use/tool_result 配对（单遍正向扫描）
     4. 剥离 thinking blocks（signature 无效）
+    5. 独立的 Anthropic 兼容性自检（纯检测，不修改，定位 enforce/sanity 未覆盖的边界 case）
 
     所有变换均为幂等操作，安全地在已清理的请求体上重复执行。
 
@@ -686,6 +798,10 @@ def prepare_zhipu_to_anthropic(
     """
     prepared = copy.deepcopy(body)
     adaptations: list[str] = []
+    msgs = prepared.get("messages", [])
+
+    # ── 过渡管线诊断：变换前快照 ──
+    _dump_message_digest(msgs, label="zhipu→anthropic.before")
 
     # Step 1: 剥离 zhipu 私有 content block 类型（如 server_tool_use_delta）
     removed_vendor_blocks = _remove_vendor_blocks(prepared, _ZHIPU_VENDOR_BLOCK_TYPES)
@@ -696,16 +812,24 @@ def prepare_zhipu_to_anthropic(
     rewritten, _ = _rewrite_srvtoolu_ids(prepared)
     if rewritten:
         adaptations.append(f"rewritten_{rewritten}_srvtoolu_ids")
+    _dump_message_digest(msgs, label="zhipu→anthropic.after_rewrite")
 
     # Step 3: 强制 tool_use/tool_result 配对
-    pairing_fixes = enforce_anthropic_tool_pairing(prepared.get("messages", []))
+    pairing_fixes = enforce_anthropic_tool_pairing(msgs)
     if pairing_fixes:
         adaptations.extend(pairing_fixes)
+    _dump_message_digest(msgs, label="zhipu→anthropic.after_enforce")
 
     # Step 4: 剥离 thinking blocks（zhipu signature 无效）
     stripped = strip_thinking_blocks(prepared)
     if stripped:
         adaptations.append(f"stripped_{stripped}_thinking_blocks")
+    _dump_message_digest(msgs, label="zhipu→anthropic.after_strip")
+
+    # Step 5: 独立的 Anthropic 兼容性自检（纯检测，不修改）
+    validation_issues = _validate_anthropic_pairing(msgs, context="zhipu→anthropic")
+    if validation_issues:
+        adaptations.append("anthropic_pairing_validation_issues")
 
     return prepared, adaptations
 

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/src/coding/proxy/server/dashboard.py

@@ -89,6 +89,7 @@ _DASHBOARD_HTML = """<!DOCTYPE html>
       --shadow-md: 0 8px 24px rgba(0,0,0,.3);
       --glow-blue: 0 0 0 1px rgba(88,166,255,.1), 0 8px 32px rgba(88,166,255,.04);
       --gradient-primary: linear-gradient(135deg, #667eea, #764ba2);
+      --gap-section: 12px;
     }
     @keyframes fadeInUp {
       from { opacity: 0; transform: translateY(10px); }
@@ -160,7 +161,7 @@ _DASHBOARD_HTML = """<!DOCTYPE html>
       display: grid;
       grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));
       gap: 5px;
-      margin-bottom: 24px;
+      margin-bottom: var(--gap-section);
     }
     .kpi-card {
       background: rgba(18,22,30,.7);
@@ -214,13 +215,13 @@ _DASHBOARD_HTML = """<!DOCTYPE html>
       display: grid;
       grid-template-columns: 1fr 2fr;
       gap: 16px;
-      margin-bottom: 16px;
+      margin-bottom: var(--gap-section);
     }
     .charts-grid-2 {
       display: grid;
       grid-template-columns: 1fr 2fr;
       gap: 16px;
-      margin-bottom: 16px;
+      margin-bottom: var(--gap-section);
     }
     .charts-grid > .card,
     .charts-grid-2 > .card {
@@ -356,7 +357,7 @@ _DASHBOARD_HTML = """<!DOCTYPE html>
     /* ── 时间区间选择栏 ── */
     .time-range-bar {
       display: flex; align-items: center; gap: 8px;
-      margin-bottom: 24px; flex-wrap: wrap;
+      margin-bottom: var(--gap-section); flex-wrap: wrap;
       padding: 8px 16px;
       background: rgba(18,22,30,.5);
       border: 1px solid rgba(255,255,255,.04);
@@ -560,7 +561,10 @@ _DASHBOARD_HTML = """<!DOCTYPE html>
 
     /* ── Model Calling 实时状态 ────────────────────────── */
     .model-calling-card {
-      margin-bottom: 5px;
+      margin-bottom: var(--gap-section);
+    }
+    .model-token-card {
+      margin-bottom: var(--gap-section);
     }
     .mc-empty {
       text-align: center;
@@ -791,7 +795,7 @@ _DASHBOARD_HTML = """<!DOCTYPE html>
   </div>
 
   <!-- Token 用量（按 Vendor / 模型）堆叠图 -->
-  <div class="card" style="margin-bottom:12px">
+  <div class="card model-token-card">
     <div class="card-title" id="title-model-token-timeline">近 7 天 Token 用量（按 Vendor / 模型）</div>
     <div class="chart-with-legend">
       <div class="chart-wrap-xl">
@@ -1346,7 +1350,7 @@ function startModelCallingPoll() {
     }).catch(function() {});
   }
   tick();
-  _mcTimer = setInterval(tick, 1500);
+  _mcTimer = setInterval(tick, 10000);
 }
 function stopModelCallingPoll() {
   if (_mcTimer) { clearInterval(_mcTimer); _mcTimer = null; }

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/src/coding/proxy/server/routes.py

@@ -8,7 +8,7 @@ from typing import Any
 
 import httpx
 from fastapi import Request, Response
-from fastapi.responses import StreamingResponse
+from fastapi.responses import RedirectResponse, StreamingResponse
 
 from ..vendors.base import NoCompatibleVendorError
 
@@ -197,11 +197,15 @@ def register_health_routes(app: Any) -> None:
         return {"status": "ok"}
 
     @app.head("/")
-    @app.get("/")
-    async def root() -> Response:
+    async def root_head() -> Response:
         """根路径连通性探测 — Claude Code 在建连前发送 HEAD / 作为 health probe."""
         return Response(status_code=200)
 
+    @app.get("/")
+    async def root_get() -> RedirectResponse:
+        """GET / 重定向到 Dashboard."""
+        return RedirectResponse(url="/dashboard", status_code=307)
+
 
 def register_status_route(app: Any, router: Any) -> None:
     """注册状态查询路由."""

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/tests/e2e/test_e2e_http.py

@@ -188,13 +188,16 @@ async def test_http_health_probe(e2e_client: object) -> None:
     )
 
     get_resp = await e2e_client.get("/")
-    assert get_resp.status_code == 200, f"GET / 预期 200，实际 {get_resp.status_code}"
+    assert get_resp.status_code == 307, f"GET / 预期 307，实际 {get_resp.status_code}"
+    assert get_resp.headers["location"] == "/dashboard"
 
     health_resp = await e2e_client.get("/health")
     assert health_resp.status_code == 200
     assert health_resp.json() == {"status": "ok"}
 
-    print("\n[E2E] HTTP health probe 成功: HEAD /=200, GET /=200, /health=ok")
+    print(
+        "\n[E2E] HTTP health probe 成功: HEAD /=200, GET /=307→/dashboard, /health=ok"
+    )
 
 
 @pytest.mark.e2e

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/tests/test_app_routes.py

@@ -35,11 +35,12 @@ def test_head_root_returns_200():
         assert resp.status_code == 200
 
 
-def test_get_root_returns_200():
-    """GET / 返回 200."""
+def test_get_root_redirects_to_dashboard():
+    """GET / 重定向到 /dashboard."""
     with _make_app() as client:
-        resp = client.get("/")
-        assert resp.status_code == 200
+        resp = client.get("/", follow_redirects=False)
+        assert resp.status_code == 307
+        assert resp.headers["location"] == "/dashboard"
 
 
 # ── count_tokens 透传 ────────────────────────────────────────

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/tests/test_vendor_channels.py

@@ -2225,3 +2225,221 @@ class TestNormalizeForZhipu:
         assert result["stream"] is True
         assert result["metadata"] == {"user_id": "test"}
         assert adaptations == []
+
+
+class TestDumpMessageDigest:
+    """``_dump_message_digest`` 诊断快照函数测试."""
+
+    def test_outputs_nothing_on_empty_messages(self, caplog):
+        import logging
+
+        from coding.proxy.convert.vendor_channels import _dump_message_digest
+
+        with caplog.at_level(
+            logging.DEBUG, logger="coding.proxy.convert.vendor_channels"
+        ):
+            _dump_message_digest([], label="test")
+        assert "Transition digest" not in caplog.text
+
+    def test_outputs_structure_for_first_n_messages(self, caplog):
+        import logging
+
+        from coding.proxy.convert.vendor_channels import _dump_message_digest
+
+        messages = [
+            {"role": "user", "content": [{"type": "text", "text": "hi"}]},
+            {
+                "role": "assistant",
+                "content": [
+                    {"type": "tool_use", "id": "toolu_1", "name": "bash", "input": {}},
+                ],
+            },
+            {
+                "role": "user",
+                "content": [
+                    {"type": "tool_result", "tool_use_id": "toolu_1", "content": "ok"},
+                ],
+            },
+        ]
+        with caplog.at_level(
+            logging.DEBUG, logger="coding.proxy.convert.vendor_channels"
+        ):
+            _dump_message_digest(messages, label="test")
+        assert "test" in caplog.text
+        assert "0:user" in caplog.text
+        assert "1:assistant" in caplog.text
+        assert "tool_use:1" in caplog.text
+
+
+class TestValidateAnthropicPairing:
+    """``_validate_anthropic_pairing`` 独立配对自检测试."""
+
+    def test_no_issues_for_correct_pairing(self):
+        from coding.proxy.convert.vendor_channels import _validate_anthropic_pairing
+
+        messages = [
+            {"role": "user", "content": "go"},
+            {
+                "role": "assistant",
+                "content": [
+                    {"type": "tool_use", "id": "toolu_1", "name": "bash", "input": {}},
+                ],
+            },
+            {
+                "role": "user",
+                "content": [
+                    {"type": "tool_result", "tool_use_id": "toolu_1", "content": "ok"},
+                ],
+            },
+        ]
+        issues = _validate_anthropic_pairing(messages)
+        assert issues == []
+
+    def test_detects_missing_tool_result(self):
+        from coding.proxy.convert.vendor_channels import _validate_anthropic_pairing
+
+        messages = [
+            {"role": "user", "content": "go"},
+            {
+                "role": "assistant",
+                "content": [
+                    {"type": "tool_use", "id": "toolu_1", "name": "bash", "input": {}},
+                ],
+            },
+            {"role": "user", "content": [{"type": "text", "text": "no result"}]},
+        ]
+        issues = _validate_anthropic_pairing(messages)
+        assert len(issues) == 1
+        assert "toolu_1" in issues[0]
+        assert "messages[1]" in issues[0]
+
+    def test_detects_non_user_after_assistant_with_tool_use(self):
+        from coding.proxy.convert.vendor_channels import _validate_anthropic_pairing
+
+        messages = [
+            {"role": "user", "content": "go"},
+            {
+                "role": "assistant",
+                "content": [
+                    {"type": "tool_use", "id": "toolu_1", "name": "bash", "input": {}},
+                ],
+            },
+            {
+                "role": "assistant",
+                "content": [{"type": "text", "text": "another assistant"}],
+            },
+        ]
+        issues = _validate_anthropic_pairing(messages)
+        assert len(issues) == 1
+        assert "not user" in issues[0]
+
+    def test_detects_assistant_with_tool_use_at_end_of_list(self):
+        from coding.proxy.convert.vendor_channels import _validate_anthropic_pairing
+
+        messages = [
+            {"role": "user", "content": "go"},
+            {
+                "role": "assistant",
+                "content": [
+                    {"type": "tool_use", "id": "toolu_1", "name": "bash", "input": {}},
+                ],
+            },
+        ]
+        issues = _validate_anthropic_pairing(messages)
+        assert len(issues) == 1
+        assert "end of list" in issues[0]
+
+    def test_partial_missing_only_reports_missing_ids(self):
+        from coding.proxy.convert.vendor_channels import _validate_anthropic_pairing
+
+        messages = [
+            {"role": "user", "content": "go"},
+            {
+                "role": "assistant",
+                "content": [
+                    {"type": "tool_use", "id": "toolu_1", "name": "bash", "input": {}},
+                    {"type": "tool_use", "id": "toolu_2", "name": "read", "input": {}},
+                ],
+            },
+            {
+                "role": "user",
+                "content": [
+                    {"type": "tool_result", "tool_use_id": "toolu_1", "content": "ok"},
+                ],
+            },
+        ]
+        issues = _validate_anthropic_pairing(messages)
+        assert len(issues) == 1
+        assert "toolu_2" in issues[0]
+        assert "1/2" in issues[0]
+
+    def test_integration_with_zhipu_to_anthropic_channel(self):
+        """验证 prepare_zhipu_to_anthropic 在末端执行自检且 adaptations 包含标签."""
+        from coding.proxy.convert.vendor_channels import prepare_zhipu_to_anthropic
+
+        body = {
+            "model": "claude-opus-4-7",
+            "messages": [
+                {"role": "user", "content": "go"},
+                {
+                    "role": "assistant",
+                    "content": [
+                        {
+                            "type": "server_tool_use",
+                            "id": "srvtoolu_01",
+                            "name": "bash",
+                            "input": {},
+                        },
+                    ],
+                },
+                {
+                    "role": "user",
+                    "content": [
+                        {
+                            "type": "tool_result",
+                            "tool_use_id": "srvtoolu_01",
+                            "content": "ok",
+                        },
+                    ],
+                },
+            ],
+        }
+        result, adaptations = prepare_zhipu_to_anthropic(body)
+        # 自检通过，不应包含 validation_issues 标签
+        assert "anthropic_pairing_validation_issues" not in adaptations
+
+    def test_integration_detects_enforce_missed_issue(self):
+        """构造一个理论上 enforce 可能遗漏的场景，验证自检能捕获.
+
+        场景：两条连续 assistant 消息，第一条的 tool_result 被第二条的
+        existing_result_ids"冒领"（相同 ID 碰撞场景的模拟）。
+        虽然当前 enforce 实现下不太可能自然产生此场景，但自检应能捕获。
+        """
+        from coding.proxy.convert.vendor_channels import (
+            _validate_anthropic_pairing,
+        )
+
+        # 手动构造一个 enforce 后仍存在配对缺陷的 body
+        messages = [
+            {"role": "user", "content": "go"},
+            {
+                "role": "assistant",
+                "content": [
+                    {"type": "tool_use", "id": "toolu_x", "name": "bash", "input": {}},
+                ],
+            },
+            {
+                "role": "user",
+                "content": [
+                    # tool_result 缺失 tolu_x，但有不相关的 tool_result
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_other",
+                        "content": "wrong",
+                    },
+                ],
+            },
+        ]
+        issues = _validate_anthropic_pairing(messages)
+        assert len(issues) == 1
+        assert "toolu_x" in issues[0]

{coding_proxy-0.5.1a1 → coding_proxy-0.5.1a3}/uv.lock

@@ -74,7 +74,7 @@ wheels = [
 
 [[package]]
 name = "coding-proxy"
-version = "0.5.1a1"
+version = "0.5.1a3"
 source = { editable = "." }
 dependencies = [
     { name = "aiosqlite" },