docs(wecom): add config set commands for text templates

Co-authored-by: Cursor <cursoragent@cursor.com>

Author: hiwepy

extensions/wecom/README.en.md

@@ -558,6 +558,49 @@ JSON does not allow comments. Keys are ordered by role: **welcome & stream proto
 }
 ```
 
+### Per-Key Command Setup
+
+Use `channels.wecom.<key>` for global default copy. In multi-account setups, replace the prefix with `channels.wecom.accounts.<accountId>.<key>` for account-level overrides, for example `channels.wecom.accounts.cs-assistant.thinkingText`. The Agent app currently supports a separate welcome override only: `channels.wecom.agent.welcomeText`; other runtime text templates still use the global or account-level flat `*Text` fields.
+
+The sample values below use the Chinese copy from the JSON example; replace them with localized copy as needed. Keep placeholders such as `{toolName}`, `{elapsed}`, `{minutes}`, `{kind}`, `{detail}`, `{mediaUrl}`, `{reason}`, and `{emptyReply}` unchanged because runtime code replaces them.
+
+```bash
+openclaw config set channels.wecom.welcomeText "您好！我是智能助手，发送消息即可开始对话。"
+openclaw config set channels.wecom.streamPlaceholderText "1"
+openclaw config set channels.wecom.thinkingText "🤔 正在思考…"
+openclaw config set channels.wecom.receivedText "📩 已收到…"
+openclaw config set channels.wecom.toolStatusText "正在调用 {toolName}…"
+openclaw config set channels.wecom.readingText "📎 正在阅读附件…"
+openclaw config set channels.wecom.generatingText "✍️ 正在输入…"
+openclaw config set channels.wecom.compactionText "📦 正在压缩…"
+openclaw config set channels.wecom.emptyReplyText "⚠️ 未能生成可展示的回复，请稍后重试或发送文字消息。"
+openclaw config set channels.wecom.finishFooterText "⏱ {elapsed}s · 已完成"
+openclaw config set channels.wecom.cardSentText "📋 卡片消息已发送。"
+openclaw config set channels.wecom.mediaSentText "📎 文件已发送，请查收。"
+openclaw config set channels.wecom.mediaParseFailedText "⚠️ 未能解析该媒体并生成回复。{emptyReply}"
+openclaw config set channels.wecom.mediaDeliveredText "✅ 文件已发送。"
+openclaw config set channels.wecom.processedCompleteText "✅ 已处理完成。"
+openclaw config set channels.wecom.timeoutText "⚠️ 处理超时（约 {minutes} 分钟），请稍后重试或发送文字消息。"
+openclaw config set channels.wecom.dispatchErrorText "⚠️ 回复生成失败（{kind}）：{detail}"
+openclaw config set channels.wecom.mediaErrorNoAccessText $'⚠️ 文件发送失败：没有权限访问路径 {mediaUrl}\n请在 openclaw.json 的 mediaLocalRoots 中添加该路径的父目录后重启生效。'
+openclaw config set channels.wecom.mediaErrorReasonText "⚠️ 文件发送失败：{reason}"
+openclaw config set channels.wecom.mediaErrorGenericText "⚠️ 文件发送失败：无法处理文件 {mediaUrl}，请稍后再试。"
+openclaw config set channels.wecom.queuedText "⏳ 排队中…"
+openclaw config set channels.wecom.mergedQueuedText "⏳ 已合并排队…"
+openclaw config set channels.wecom.mergedDoneText "✅ 已合并处理完成，请查看上一条回复。"
+openclaw config set channels.wecom.sessionResetText "✅ 已重置会话。"
+openclaw config set channels.wecom.sessionNewText "✅ 已开启新会话。"
+openclaw gateway restart
+```
+
+`mediaErrorNoAccessText` contains a newline. zsh/bash users should use ANSI-C quoting as shown above (`$'...\n...'`). If your shell does not support it, write the value through JSON/file-based config instead so `\n` is stored as an actual newline.
+
+For an Agent app welcome message, set the dedicated override:
+
+```bash
+openclaw config set channels.wecom.agent.welcomeText "欢迎使用自建应用，我会尽快回复您。"
+```
+
 ## Streaming Output
 
 Only **Bot WebSocket** and **Bot Webhook** support WeCom `stream` / `replyStream`. **Agent inbound chat does not use Bot-style streaming**; outbound delivery is primarily one-shot Markdown / media via the Agent API.

extensions/wecom/README.md

@@ -562,6 +562,49 @@ JSON 不支持注释；下方按职责分组排列：**欢迎与流式协议** 
 }
 ```
 
+### 逐项命令设置
+
+全局默认文案使用 `channels.wecom.<key>`。多账号场景下，账号级覆盖可将前缀替换为 `channels.wecom.accounts.<accountId>.<key>`，例如 `channels.wecom.accounts.cs-assistant.thinkingText`。自建应用当前仅支持独立覆盖欢迎语：`channels.wecom.agent.welcomeText`；其他运行中文案仍使用全局或账号级平铺 `*Text` 字段。
+
+下面命令可直接复制执行；包含 `{toolName}`、`{elapsed}`、`{minutes}`、`{kind}`、`{detail}`、`{mediaUrl}`、`{reason}`、`{emptyReply}` 的占位符需要原样保留，由运行时替换。
+
+```bash
+openclaw config set channels.wecom.welcomeText "您好！我是智能助手，发送消息即可开始对话。"
+openclaw config set channels.wecom.streamPlaceholderText "1"
+openclaw config set channels.wecom.thinkingText "🤔 正在思考…"
+openclaw config set channels.wecom.receivedText "📩 已收到…"
+openclaw config set channels.wecom.toolStatusText "正在调用 {toolName}…"
+openclaw config set channels.wecom.readingText "📎 正在阅读附件…"
+openclaw config set channels.wecom.generatingText "✍️ 正在输入…"
+openclaw config set channels.wecom.compactionText "📦 正在压缩…"
+openclaw config set channels.wecom.emptyReplyText "⚠️ 未能生成可展示的回复，请稍后重试或发送文字消息。"
+openclaw config set channels.wecom.finishFooterText "⏱ {elapsed}s · 已完成"
+openclaw config set channels.wecom.cardSentText "📋 卡片消息已发送。"
+openclaw config set channels.wecom.mediaSentText "📎 文件已发送，请查收。"
+openclaw config set channels.wecom.mediaParseFailedText "⚠️ 未能解析该媒体并生成回复。{emptyReply}"
+openclaw config set channels.wecom.mediaDeliveredText "✅ 文件已发送。"
+openclaw config set channels.wecom.processedCompleteText "✅ 已处理完成。"
+openclaw config set channels.wecom.timeoutText "⚠️ 处理超时（约 {minutes} 分钟），请稍后重试或发送文字消息。"
+openclaw config set channels.wecom.dispatchErrorText "⚠️ 回复生成失败（{kind}）：{detail}"
+openclaw config set channels.wecom.mediaErrorNoAccessText $'⚠️ 文件发送失败：没有权限访问路径 {mediaUrl}\n请在 openclaw.json 的 mediaLocalRoots 中添加该路径的父目录后重启生效。'
+openclaw config set channels.wecom.mediaErrorReasonText "⚠️ 文件发送失败：{reason}"
+openclaw config set channels.wecom.mediaErrorGenericText "⚠️ 文件发送失败：无法处理文件 {mediaUrl}，请稍后再试。"
+openclaw config set channels.wecom.queuedText "⏳ 排队中…"
+openclaw config set channels.wecom.mergedQueuedText "⏳ 已合并排队…"
+openclaw config set channels.wecom.mergedDoneText "✅ 已合并处理完成，请查看上一条回复。"
+openclaw config set channels.wecom.sessionResetText "✅ 已重置会话。"
+openclaw config set channels.wecom.sessionNewText "✅ 已开启新会话。"
+openclaw gateway restart
+```
+
+`mediaErrorNoAccessText` 含换行，zsh/bash 推荐使用上面的 ANSI-C quoting（`$'...\n...'`）。如果你的 shell 不支持该写法，请改用 JSON/文件方式写入配置，避免把 `\n` 写成普通文本。
+
+自建应用欢迎语可单独覆盖：
+
+```bash
+openclaw config set channels.wecom.agent.welcomeText "欢迎使用自建应用，我会尽快回复您。"
+```
+
 ## 流式输出
 
 仅 **Bot WebSocket** 与 **Bot Webhook** 支持企业微信 `stream` / `replyStream` 流式载体；**Agent 自建应用入站对话不支持 Bot 式流式**，出站以 `sendMessage` 一次性 Markdown / 媒体为主。