Fix Bedrock tool history and add upload limit config

scripts/configure-nginx-upload-limit.sh

@@ -0,0 +1,28 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# 配置线上 nginx 上传体积限制。
+# 需要 sudo 权限；应用服务本身仍然用普通用户运行。
+
+LIMIT="${CLAW_NGINX_CLIENT_MAX_BODY_SIZE:-300m}"
+CONF_PATH="${CLAW_NGINX_UPLOAD_CONF:-/etc/nginx/conf.d/zk-data-agent-upload.conf}"
+
+if ! command -v nginx >/dev/null 2>&1; then
+	echo "WARN 未发现 nginx，跳过上传限制配置。"
+	exit 0
+fi
+
+if ! command -v sudo >/dev/null 2>&1; then
+	echo "ERROR 配置 nginx 需要 sudo。" >&2
+	exit 1
+fi
+
+sudo tee "${CONF_PATH}" >/dev/null <<EOF
+# Managed by zk-data-agent.
+# 放开 Web 文件上传体积，避免大文件上传被 nginx 直接 413。
+client_max_body_size ${LIMIT};
+EOF
+
+sudo nginx -t
+sudo systemctl reload nginx
+echo "已配置 nginx client_max_body_size=${LIMIT} (${CONF_PATH})"

scripts/deploy-ubuntu.sh

@@ -122,6 +122,17 @@ bootstrap_system_packages() {
 		systemd
 }
 
+configure_nginx_upload_limit() {
+	if [[ "${BOOTSTRAP_SYSTEM}" != "1" && -f "${ENV_FILE}" ]]; then
+		return
+	fi
+	local script="${ROOT_DIR}/scripts/configure-nginx-upload-limit.sh"
+	if [[ -x "${script}" ]]; then
+		log "配置 nginx 上传体积限制"
+		"${script}" || warn "nginx 上传体积限制配置失败，可稍后手动执行：bash ${script}"
+	fi
+}
+
 prompt_value() {
 	local var_name="$1"
 	local prompt="$2"
@@ -345,6 +356,7 @@ health_check() {
 main() {
 	log "ZK Data Agent Ubuntu 部署"
 	bootstrap_system_packages
+	configure_nginx_upload_limit
 	update_git
 	ensure_env_file
 	configure_instance_names

src/openai_compat.py

@@ -91,6 +91,99 @@ def _parse_tool_arguments(raw_arguments: Any) -> dict[str, Any]:
     )
 
 
+def _tool_call_ids(message: dict[str, Any]) -> list[str]:
+    raw_tool_calls = message.get('tool_calls')
+    if not isinstance(raw_tool_calls, list):
+        return []
+    ids: list[str] = []
+    for index, raw_call in enumerate(raw_tool_calls):
+        if not isinstance(raw_call, dict):
+            continue
+        call_id = raw_call.get('id')
+        ids.append(call_id if isinstance(call_id, str) and call_id else f'call_{index}')
+    return ids
+
+
+def _contiguous_tool_result_ids(
+    messages: list[dict[str, Any]],
+    start_index: int,
+) -> set[str]:
+    ids: set[str] = set()
+    index = start_index
+    while index < len(messages) and messages[index].get('role') == 'tool':
+        tool_call_id = messages[index].get('tool_call_id')
+        if isinstance(tool_call_id, str) and tool_call_id:
+            ids.add(tool_call_id)
+        index += 1
+    return ids
+
+
+def _filter_tool_calls_by_ids(
+    message: dict[str, Any],
+    valid_ids: set[str],
+) -> dict[str, Any]:
+    raw_tool_calls = message.get('tool_calls')
+    if not isinstance(raw_tool_calls, list):
+        return message
+    filtered: list[dict[str, Any]] = []
+    for index, raw_call in enumerate(raw_tool_calls):
+        if not isinstance(raw_call, dict):
+            continue
+        call_id = raw_call.get('id')
+        normalized_id = (
+            call_id if isinstance(call_id, str) and call_id else f'call_{index}'
+        )
+        if normalized_id in valid_ids:
+            filtered.append(raw_call)
+    if len(filtered) == len(raw_tool_calls):
+        return message
+    next_message = dict(message)
+    if filtered:
+        next_message['tool_calls'] = filtered
+    else:
+        next_message.pop('tool_calls', None)
+    return next_message
+
+
+def _sanitize_anthropic_tool_messages(
+    messages: list[dict[str, Any]],
+) -> list[dict[str, Any]]:
+    """清理 Anthropic/Bedrock 严格要求的 tool_use/tool_result 邻接关系。
+
+    OpenAI 风格历史在异常中断、裁剪或前端恢复时可能留下 assistant tool_calls，
+    但下一条消息不再紧跟对应 tool 结果。Anthropic Messages API 会直接拒绝
+    这种历史。这里在出网关前删除缺失结果的 tool_call，并跳过没有前置
+    tool_use 的孤儿 tool 结果，保留普通文本上下文继续对话。
+    """
+    sanitized: list[dict[str, Any]] = []
+    pending_tool_ids: set[str] = set()
+
+    for index, message in enumerate(messages):
+        role = message.get('role')
+        if role == 'assistant':
+            tool_ids = _tool_call_ids(message)
+            if tool_ids:
+                result_ids = _contiguous_tool_result_ids(messages, index + 1)
+                valid_ids = set(tool_ids).intersection(result_ids)
+                message = _filter_tool_calls_by_ids(message, valid_ids)
+                pending_tool_ids = valid_ids
+            else:
+                pending_tool_ids = set()
+            sanitized.append(message)
+            continue
+
+        if role == 'tool':
+            tool_call_id = message.get('tool_call_id')
+            if isinstance(tool_call_id, str) and tool_call_id in pending_tool_ids:
+                sanitized.append(message)
+            continue
+
+        pending_tool_ids = set()
+        sanitized.append(message)
+
+    return sanitized
+
+
 def _optional_int(value: Any) -> int:
     if isinstance(value, bool):
         return 0
@@ -341,7 +434,7 @@ class OpenAICompatClient:
     ) -> dict[str, Any]:
         system_parts: list[str] = []
         anthropic_messages: list[dict[str, Any]] = []
-        for message in messages:
+        for message in _sanitize_anthropic_tool_messages(messages):
             role = message.get('role')
             if role == 'system':
                 content = _normalize_content(message.get('content')).strip()