dsh-data-cleaning-agent 0.8.20 → 0.8.21

package/CHANGELOG.md

@@ -4,6 +4,13 @@
 
 ## [Unreleased]
 
+## [0.8.21] - 2026-09-09
+
+- 从实际 Host 请求获取文件地址的协议和主机；修复本地 HTTP 链接被补成 HTTPS 的问题。
+- 工具卡片支持同源完整文件地址，继续拒绝非同源链接。
+- 已确认任务开始执行时自动展开工作台一次并跟随处理步骤；生成说明不触发，关闭后不反复弹出。
+- 246 项测试及 10 组浏览器回归通过，无付费查询。
+
 ## [0.8.20] - 2026-09-09
 
 - 普通无记录／未披露字段保留为无数据说明，尊重共享契约 reviewRequired=false，不再误计为待核验。

package/README.en.md

@@ -8,7 +8,7 @@ Related agents: [数据清洗补全](https://github.com/duhu2000/dsh-data-cleani
 
 > A data cleaning & completion agent plugin for DeepSeek Harness: local CSV/XLSX/JSON engine plus optional Qichacha (QCC) MCP enterprise-data enrichment. Initiated and maintained by the Qichacha (QCC) team.
 >
-> Current source version / 当前源码版本: **0.8.20** (stable release)
+> Current source version / 当前源码版本: **0.8.21** (stable release)
 
 This release unifies spreadsheet previews, column mapping, and enrichment scope across the prompt wizard and workbench. Mapping statuses use color and text; mapped fields automatically enter the scope, with optional extra fields collapsed. Multiple confirmed output columns may share a non-identity field while preserving non-empty values. It fixes saved drafts resetting the scope to five defaults and detects stale Host versions. Preparing a draft does not query QCC; sending starts execution. Fully restart DSH after upgrading. Existing download artifacts are not rewritten.
 

package/README.md

@@ -1,13 +1,13 @@
 # dsh-data-cleaning-agent
 
-本版统一向导与工作台的 Host 执行闭环：发送后自动生成新 Excel，原文件不修改，工作台入口可恢复；发布记录见 [发布说明](docs/RELEASE-0.8.20.md)。
+本版统一向导与工作台的 Host 执行闭环：发送后自动生成新 Excel，原文件不修改，工作台入口可恢复；发布记录见 [发布说明](docs/RELEASE-0.8.21.md)。
 
 ## 安装与三分钟上手
 
 数据清洗补全智能体：面向 Excel/CSV/JSON 企业名单，提供数据清洗、表格清洗、清洗补全、去重、企业数据补全与字段补全，支持企查查 MCP 和结果导出。
 
 ```sh
-dsh plugin --profile web add dsh-data-cleaning-agent@0.8.20
+dsh plugin --profile web add dsh-data-cleaning-agent@0.8.21
 ```
 
 请先满足下文的 DSH、连接器及侧边栏依赖要求；安装后完整停止并重启对应 Profile。
@@ -25,7 +25,7 @@ dsh plugin --profile web add dsh-data-cleaning-agent@0.8.20
 
 > 在 DeepSeek Harness 中清洗、补全、画像企业名单数据的智能体插件：本地 CSV/XLSX/JSON 引擎 + 可选企查查（Qichacha/QCC）MCP 企业数据补全，由企查查（Qichacha/QCC）团队发起并维护。
 >
-> 当前源码版本 / Current source version: **0.8.20**（正式版本）
+> 当前源码版本 / Current source version: **0.8.21**（正式版本）
 
 本版上传表格后展示完整清单及空白数量，用绿色、琥珀色、红色及文字区分已通过、待确认和未匹配映射。已映射字段自动纳入补全范围，额外字段默认折叠；多个原列可经确认使用同一非主体字段，分别补空并保留非空原值。修复 Host 草稿重置为默认五字段的问题，增加前后端版本错配提示。生成说明不查询，发送才执行；升级后需完整重启 DSH，历史下载文件不自动改写。
 

package/docs/RELEASE-0.8.21.md

@@ -0,0 +1,9 @@
+# 0.8.21 发布说明
+
+文件预览与下载地址使用实际请求的协议、主机，不硬编码 HTTPS 或端口；工具结果卡片接受同源完整地址。
+
+检测已暂存任务开始执行后自动展开右侧工作台一次，切到当前处理步骤。生成说明不会提前打开，用户关闭后同一监听不会重复打开。
+
+验证：246 项测试及 10 组浏览器回归。未重新执行付费查询，未修改上传文件。
+
+升级后完整重启 DSH，并重新生成后续任务说明。历史对话中的错误 HTTPS 文本链接不会被改写，可通过工作台打开历史文件。

package/lib/client.js

@@ -1862,9 +1862,14 @@ body:has(.dcAgentPromptBackdrop) { overflow: hidden; }
       return response.task;
     }
     function watchHostedCommand(actions, sessionId, command) {
+      let openedForExecution = false;
       void waitForQccCommand(command.commandId, 1800, async (status) => {
         if (!status || status.state === 'prepared') return;
         await refreshHostedTask(actions, sessionId, command.taskId, null, true).catch(() => {});
+        if (!openedForExecution && isCleaningSession(sessionId) && workflowTaskBySession.get(String(sessionId))?.id === command.taskId) {
+          openedForExecution = true;
+          openWorkbench(actions, workflowTaskBySession.get(String(sessionId))?.stage || 'match', sessionId);
+        }
         if (isCleaningSession(sessionId) && workflowTaskBySession.get(String(sessionId))?.id === command.taskId) actions.setCommandProgress?.({ taskId: command.taskId, ...status.progress });
       }).then(async (completed) => {
         await refreshHostedTask(actions, sessionId, command.taskId, completed.run);
@@ -3228,7 +3233,7 @@ body:has(.dcAgentPromptBackdrop) { overflow: hidden; }
         meta.hint ? h('div', { className: 'dcAgentToolCardHint' }, meta.hint) : null,
         summary ? h('div', { className: 'dcAgentToolCardBody' }, summary) : null,
         toolName === 'data_cleaning_qcc_run' ? h('div', { className: 'dcAgentArtifactList' },
-          [...summary.matchAll(/\[([^\]]+)\]\((\/data-cleaning\/api\/workflow\/tasks\/[a-zA-Z0-9%_-]+\/artifacts\/[a-zA-Z0-9%_-]+(?:\?preview=1)?)\)/g)].map((match, index) =>
+          [...summary.matchAll(/\[([^\]]+)\]\(((?:https?:\/\/[a-zA-Z0-9.:[\]-]+)?\/data-cleaning\/api\/workflow\/tasks\/[a-zA-Z0-9%_-]+\/artifacts\/[a-zA-Z0-9%_-]+(?:\?preview=1)?)\)/g)].filter(match => match[2].startsWith('/') || (typeof location !== 'undefined' && new URL(match[2]).origin === location.origin)).map((match, index) =>
             h('a', { key: index, className: 'dcAgentButton', href: match[2], target: '_blank', rel: 'noopener noreferrer' }, match[1]))) : null,
       );
     }

package/lib/qcc-command.js

@@ -138,7 +138,7 @@ export class QccCommandStore {
     }
   }
 
-  prepare(input) {
+  prepare(input, { artifactOrigin = '' } = {}) {
     this.cleanup(true);
     const normalized = normalizedInput(input);
     const estimate = normalized.kind === 'enrich'
@@ -148,6 +148,7 @@ export class QccCommandStore {
     const at = new Date(this.clock()).toISOString();
     const record = {
       commandId,
+      artifactOrigin,
       taskId: normalized.taskId,
       kind: normalized.kind,
       state: 'prepared',
@@ -265,7 +266,7 @@ export class QccCommandStore {
       state: run.state,
       summary: run.summary,
       ...(record.delivery ? { deliveryState: record.delivery.state, artifactCount: record.delivery.artifacts.length,
-        artifacts: record.delivery.artifacts.filter(a => a.format === 'xlsx' && (a.kind !== 'review' || a.rowCount > 0)).map(a => ({ fileName: a.fileName, url: `/data-cleaning/api/workflow/tasks/${encodeURIComponent(record.taskId)}/artifacts/${encodeURIComponent(a.id)}` })) } : {}),
+        artifacts: record.delivery.artifacts.filter(a => a.format === 'xlsx' && (a.kind !== 'review' || a.rowCount > 0)).map(a => ({ fileName: a.fileName, url: `${record.artifactOrigin || ''}/data-cleaning/api/workflow/tasks/${encodeURIComponent(record.taskId)}/artifacts/${encodeURIComponent(a.id)}` })) } : {}),
     });
   }
 }

package/lib/web.js

@@ -659,11 +659,20 @@ export function mountWebRoutes(wctx, { logger, report, TOOL_NAME, SKILL_NAME })
         }
         const payload = JSON.parse((await readBody(req)).toString('utf8'));
         if (payload.kind && payload.kind !== 'enrich') requirePaidConfirmation(payload);
+        // Resolve from the actual browser request, never from model/task input.
+        let artifactOrigin = '';
+        const requestHost = req.headers?.host;
+        if (typeof requestHost === 'string' && /^[a-zA-Z0-9.:[\]-]+$/.test(requestHost)) {
+          try {
+            const origin = new URL(req.headers.origin || `${req.socket?.encrypted ? 'https' : 'http'}://${requestHost}`);
+            if (['http:', 'https:'].includes(origin.protocol) && origin.host === requestHost) artifactOrigin = origin.origin;
+          } catch { /* Relative links remain available for hosts without an origin. */ }
+        }
         return writeJson(res, 201, {
           ok: true,
           marker: 'g5-agent-command',
           paidCalls: false,
-          command: qccCommands.prepare(payload.workflowOwned ? await execution.prepare(payload) : payload),
+          command: qccCommands.prepare(payload.workflowOwned ? await execution.prepare(payload) : payload, { artifactOrigin }),
         });
       }
       if (rest.length === 1 && req.method === 'GET') {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "dsh-data-cleaning-agent",
-  "version": "0.8.20",
+  "version": "0.8.21",
   "description": "Data cleaning and data enrichment for CSV/XLSX/JSON enterprise lists in DeepSeek Harness, including spreadsheet cleaning, deduplication, profiling, optional Qichacha MCP and exports.",
   "type": "module",
   "main": "lib/index.js",
@@ -40,7 +40,7 @@
     "docs/RELEASE-0.8.1.md",
     "docs/RELEASE-0.8.16.md",
     "docs/RELEASE-0.8.17.md",
-    "docs/RELEASE-0.8.20.md",
+    "docs/RELEASE-0.8.21.md",
     "docs/TEMPLATE-MAPPING-ACCEPTANCE.md",
     "docs/UI-WORKFLOW-V2.md",
     "docs/UI-WORKFLOW-V2-MIGRATION.md",

package/docs/RELEASE-0.8.20.md

@@ -1,11 +0,0 @@
-# 0.8.20 发布说明
-
-修复未披露数据被误判为人工核验；保留未披露说明和空白值，不伪造数据。真正的多候选或无法判定的问题继续进入待核验。
-
-当前页和历史列表同步最新任务状态。处理进度条区分处理中、完成、需核验和失败；100% 仅表示主体处理结束。摘要尚未生成时显示“结果统计中”。
-
-任务展示名优先尊重自定义名称，否则由源文件名、行数和创建时间生成；当前页与历史一致。
-
-验证：246 项自动测试、10 组浏览器回归。无付费查询，源文件未修改。
-
-升级并完整重启 DSH 后生效；历史任务的旧异常文件和统计不会自动重算，不应仅为更新展示而重复查询。