全栈安全加固、性能提速与架构演进:连接池与沙箱止血、双路混合检索、多版本归档及前端分包治理

- 第一批(安全加固与生命周期止血):隔离测试宿主机密钥与 .env 泄漏;修复 SQLite 取消查询挂死、:memory: 连接池报错与 EngineCache 世代竞态;补齐运行与评测删除时的 Checkpointer 线程、Artifact 磁盘文件与无主附件回收;加固三套沙箱路径越界、进程组清理与 Seatbelt 数据目录隔离,封堵 SSRF IPv4-mapped IPv6 元数据绕过与响应体 OOM;修复推理模型多轮 reasoning_content、结构化输出 Token 漏计与缓存折扣计费
- 第二批(数据库/集群/问数提速与首屏分包):补齐核心表索引与 SQLite WAL 调优,列表接口按需 defer 大列,收敛启动扫描并打通跨进程 slot 槽位唤醒;优化跨源全局选表、哈希桶数字绑定、流式 CSV 导入及多方言分析算子;剥离首屏静态引入的 @xyflow/react 与全局样式(首屏主包降至 461.59 kB),补充流式 rAF 攒批、Store 精细订阅与表单无障碍绑定
- 第三批(架构演进与工程自动化闭环):记录库超挂接阈值历史版本自动增量归档至单库多表(解除 SQLite 10 库 ATTACH 硬上限);重构知识库为词面 BM25 与全集合归一化向量矩阵真·双路独立召回并支持游标分批重建;升级多方言感知 SQL 模板词法分析(E'...' / $$...$$ / q'[...]' / # / 嵌套注释)与 Decimal / bool / NaN 参数校验,消除 ledger 层对 api 路由层的反向依赖;拆分 3101 行 CaliberDrafts 页面为 7 个领域子模块,消除检查脚本本机路径硬编码并新增 GitHub Actions CI 门禁与三批回归测试套件
This commit is contained in:
ThearchyHelios
2026-11-13 14:51:36 +08:00
parent 09949fc8ac
commit eddeded2a4
106 changed files with 7178 additions and 3913 deletions
+70 -33
View File
@@ -21,10 +21,19 @@ _BLOCKED_PORTS = {22, 23, 25, 135, 139, 445, 3389}
_METADATA_HOSTS = frozenset({"100.100.100.200", "fd00:ec2::254", "metadata.google.internal"})
def _is_private(ip: str) -> bool:
def _unwrap_ip(ip: str) -> ipaddress.IPv4Address | ipaddress.IPv6Address | None:
try:
addr = ipaddress.ip_address(ip)
except ValueError:
return None
if isinstance(addr, ipaddress.IPv6Address) and addr.ipv4_mapped is not None:
return addr.ipv4_mapped
return addr
def _is_private(ip: str) -> bool:
addr = _unwrap_ip(ip)
if addr is None:
return True
return (
addr.is_private
@@ -48,9 +57,8 @@ def _fake_networks() -> list[ipaddress.IPv4Network | ipaddress.IPv6Network]:
def _in(ip: str, networks) -> bool:
try:
addr = ipaddress.ip_address(ip)
except ValueError:
addr = _unwrap_ip(ip)
if addr is None:
return False
return any(addr.version == n.version and addr in n for n in networks)
@@ -101,10 +109,11 @@ async def _real_ips(host: str) -> tuple[str, ...]:
def _is_metadata(ip: str) -> bool:
try:
addr = ipaddress.ip_address(ip)
raw_addr = ipaddress.ip_address(ip)
except ValueError:
return True
return addr.is_link_local or str(addr) in _METADATA_HOSTS
addr = raw_addr.ipv4_mapped if isinstance(raw_addr, ipaddress.IPv6Address) and raw_addr.ipv4_mapped is not None else raw_addr
return raw_addr.is_link_local or addr.is_link_local or str(raw_addr) in _METADATA_HOSTS or str(addr) in _METADATA_HOSTS
async def assert_safe_url(url: str, *, allow_intranet: bool = False) -> str:
@@ -199,35 +208,63 @@ async def safe_request(
max_bytes = max_bytes or settings.http_tool_max_bytes
current = await assert_safe_url(url)
_SENSITIVE_HEADERS = frozenset({"authorization", "cookie", "proxy-authorization", "x-api-key"})
cur_method = method.upper()
cur_headers = dict(headers or {})
cur_params: dict[str, str] | None = params
cur_json: object | None = json_body
cur_data: str | None = data
async with httpx.AsyncClient(
timeout=timeout, follow_redirects=False, trust_env=False
) as client:
for _ in range(5):
resp = await client.request(
method.upper(),
for hop in range(6):
async with client.stream(
cur_method,
current,
headers=headers or {},
params=params,
json=json_body,
content=data,
)
if resp.status_code in (301, 302, 303, 307, 308) and "location" in resp.headers:
current = await assert_safe_url(
str(httpx.URL(current).join(resp.headers["location"]))
)
continue
break
headers=cur_headers,
params=cur_params,
json=cur_json,
content=cur_data,
) as resp:
if hop < 5 and resp.status_code in (301, 302, 303, 307, 308) and "location" in resp.headers:
prev_url = httpx.URL(current)
next_url = prev_url.join(resp.headers["location"])
current = await assert_safe_url(str(next_url))
cur_params = None
if (next_url.scheme, next_url.host, next_url.port) != (prev_url.scheme, prev_url.host, prev_url.port):
cur_headers = {k: v for k, v in cur_headers.items() if k.lower() not in _SENSITIVE_HEADERS}
if resp.status_code == 303 or (resp.status_code in (301, 302) and cur_method == "POST"):
cur_method = "GET"
cur_json = None
cur_data = None
continue
body = resp.content[:max_bytes]
truncated = len(resp.content) > max_bytes
try:
text = body.decode(resp.encoding or "utf-8", errors="replace")
except (LookupError, UnicodeDecodeError):
text = body.decode("utf-8", errors="replace")
return {
"status": resp.status_code,
"url": str(resp.url),
"headers": dict(resp.headers),
"body": text,
"truncated": truncated,
}
chunks: list[bytes] = []
total = 0
truncated = False
async for chunk in resp.aiter_bytes():
if not chunk:
continue
remaining = max_bytes - total
if len(chunk) > remaining:
if remaining > 0:
chunks.append(chunk[:remaining])
truncated = True
break
chunks.append(chunk)
total += len(chunk)
body = b"".join(chunks)
try:
text = body.decode(resp.encoding or "utf-8", errors="replace")
except (LookupError, UnicodeDecodeError):
text = body.decode("utf-8", errors="replace")
return {
"status": resp.status_code,
"url": str(resp.url),
"headers": dict(resp.headers),
"body": text,
"truncated": truncated,
}
raise UnsafeUrlError("重定向次数超过上限")