dominds 1.9.2 → 1.9.4

package/dist/docs/memory-system.md

@@ -124,10 +124,15 @@ Taskdocs are the canonical task contract shared by the team. They answer: what w
 - `constraints`: hard rules (safety/style/process/compliance)
 - `progress`: distilled progress (key decisions, current status, next steps)
 
+#### Section semantics
+
+- `progress` is the team’s shared milestone bulletin board: key decisions, current status, and next steps. It is not a raw log and not a personal work record.
+- `goals` / `constraints` are the more stable task contract: every update must preserve still-valid entries from others instead of turning shared state into personal notes.
+
 #### Practical rules (keep only the essentials)
 
 - Treat `goals / constraints / progress` as a **shared bulletin board**, not a personal scratchpad.
-- `progress` is a distilled snapshot; high-frequency details belong in reminders (working set).
+- `progress` is a distilled snapshot; each update should preserve a complete current picture that teammates can scan quickly; high-frequency details belong in reminders (working set).
 - Don’t paste huge tool outputs into Taskdocs. Distill conclusions/evidence, or keep a curated excerpt in reminders.
 
 For extra sections and `*.tsk/` packaging semantics, see: [encapsulated-taskdoc.md](./encapsulated-taskdoc.md).
@@ -150,6 +155,8 @@ Never treat “I saw it earlier” as a durable assumption.
 Some reminders are generated by the runtime (e.g. background process status, MCP rentals/expiration). Treat them as signals:
 
 - read and adjust behavior
+- if they do not materially change your judgment/plan/risk, make no user-visible reply at all (do not send filler like “silently noted” or “received”)
+- if they do matter, reflect that only inside the next substantive reply rather than acknowledging them separately
 - but don’t treat them as your personal worklog (their lifecycle is typically system-managed)
 - in message semantics, these should be treated as **system notices**: they typically appear on the `role=user` side with an explicit notice prefix, so they do not get mistaken for self-authored work notes
 

package/dist/docs/memory-system.zh.md

@@ -123,10 +123,15 @@ TL;DR：
 - `constraints`：硬规则（安全/风格/流程/合规）
 - `progress`：精炼进展（关键决策、当前状态、下一步）
 
+#### 章节语义约定
+
+- `progress` 是全队共享的里程碑公告牌：记录关键决策、当前状态与下一步；不是流水账，不是个人工作记录。
+- `goals` / `constraints` 是较稳定的任务契约：每次更新都要保留仍然有效的他人条目，不要把共享约定误改成个人笔记。
+
 #### 使用要点（只保留最重要的约束）
 
 - 把 `goals / constraints / progress` 当作**团队共享公告板**，不是个人草稿本。
-- `progress` 只写“足以让下一程继续推进”的精炼状态；高频细节放到提醒项（工作集）。
+- `progress` 只写“足以让下一程继续推进”的精炼状态；每次更新都应保持“当前全貌快照”完整可扫读；高频细节放到提醒项（工作集）。
 - 不要把长工具输出塞进差遣牒：要么提炼成结论/证据，要么放到提醒项里做精炼摘录。
 
 差遣牒附加章节（非自动注入）与 `*.tsk/` 封装语义，详见：[encapsulated-taskdoc.zh.md](./encapsulated-taskdoc.zh.md)
@@ -149,6 +154,8 @@ TL;DR：
 有一类提醒项来自运行时环境（例如后台进程状态、MCP 租用/到期等）。它们更像信号灯：
 
 - 你应该阅读并据此调整行为
+- 若它没有实质改变当前判断/计划/风险，就禁止做任何用户可见回应（禁止写“静默吸收”“已收到”等占位语句）
+- 若它确实改变了当前判断/计划/风险，也无需单独回一条“收到”；只需在下一条有实质内容的回复中体现
 - 但它们的生命周期通常由系统管理（自动更新/自动清理），不要把它们当成个人工作日志
 - 在消息语义上，这类提醒应视为**系统提示**：通常落在 `role=user` 侧，并带显眼头标（如 `【系统提示】`），避免与个人工作提醒混淆
 

package/dist/docs/txt-editing-tools.md

@@ -114,6 +114,7 @@ Chinese version: [中文版](./txt-editing-tools.zh.md)
 
 - 每个 plan hunk 带 TTL（输出 `expires_at_ms`）。
 - hunk 存储于进程内内存；进程重启后丢失。
+- prepare 结果在 apply 前不会落盘；此时再次读取文件仍只能看到旧内容。若想基于该结果继续下一笔修改，必须先 apply 当前 hunk，再重新 prepare。
 - `apply_file_modification` 会检查：
   - hunk 是否存在且未过期
   - hunk 是否由当前成员规划（`WRONG_OWNER` 拒绝）
@@ -278,6 +279,7 @@ This is the design doc for `ws_mod` text editing as implemented.
 - Function tool calls in one message run in parallel → **prepare → apply must be two messages**.
 - Applies are serialized per file in-process (queue by `createdAtMs`, then `hunkId`).
 - `hunk_id` is TTL-limited and in-memory; apply checks ownership and access.
+- Before apply, the prepared change is not persisted, so re-reading still returns the old file content; if you want the next edit based on that prepared result, apply the current hunk first, then prepare again.
 - `create_new_file` creates a new file (empty content allowed) and refuses to overwrite existing files (YAML-only output).
 - `overwrite_entire_file` is the exception full-file overwrite tool, guarded by `known_old_total_lines/known_old_total_bytes`, and it refuses diff/patch-like content by default unless `content_format='diff'|'patch'`.
 - Some providers (e.g. Codex) may require “all fields present” in function calls; for those providers only, use sentinels like empty strings / 0 to express “unset/default”. For most providers, omit optional fields naturally.

package/dist/docs/txt-editing-tools.zh.md

@@ -114,6 +114,7 @@ updated: 2026-01-24
 
 - 每个 plan hunk 带 TTL（输出 `expires_at_ms`）。
 - hunk 存储于进程内内存；进程重启后丢失。
+- prepare 结果在 apply 前不会落盘；此时再次读取文件仍只能看到旧内容。若想基于该结果继续下一笔修改，必须先 apply 当前 hunk，再重新 prepare。
 - `apply_file_modification` 会检查：
   - hunk 是否存在且未过期
   - hunk 是否由当前成员规划（`WRONG_OWNER` 拒绝）

package/dist/llm/kernel-driver/tellask-special.js

@@ -386,15 +386,13 @@ async function finishRegisteredTellaskReplacement(args) {
     const { ownerDialog, subdialog, pendingRecord, responseBody } = args;
     const language = (0, work_language_1.getWorkLanguage)();
     const requesterId = ownerDialog.agentId;
-    const response = (0, inter_dialog_format_1.formatTellaskResponseContent)({
-        callName: pendingRecord.callName,
+    const response = (0, inter_dialog_format_1.formatTellaskReplacementNoticeContent)({
         responderId: subdialog.agentId,
         requesterId,
         mentionList: pendingRecord.mentionList,
         sessionSlug: pendingRecord.sessionSlug,
         tellaskContent: pendingRecord.tellaskContent,
         responseBody,
-        status: 'failed',
         language,
     });
     const carryoverOriginCourse = pendingRecord.callingCourse;

package/dist/minds/load.js

@@ -272,7 +272,7 @@ async function loadAgentMinds(agentId, dialog, options) {
         .filter((b) => b.trim() !== '')
         .join('\n\n');
     const intrinsicToolUsageText = (0, system_prompt_parts_1.buildIntrinsicToolUsageText)(workingLanguage, intrinsicFuncTools);
-    const toolsetManualIntro = (0, toolset_manual_1.formatToolsetManualIntro)(workingLanguage, manualTools.toolNames);
+    const toolsetManualIntro = (0, toolset_manual_1.formatToolsetManualIntro)(workingLanguage, manualTools.toolNames, toolsetNames);
     const funcToolRulesText = funcTools.length > 0 ? (0, minds_i18n_1.funcToolRulesText)(workingLanguage) : '';
     const systemPrompt = (0, system_prompt_1.buildSystemPrompt)({
         language: workingLanguage,

package/dist/minds/system-prompt-parts.js

@@ -161,9 +161,10 @@ function getMemoryPromptCopy(ctx) {
             title: '### 记忆系统（重要）',
             temporaryInfoLine: '你的聊天记录与工具输出是临时信息：会快速累积、很快过时，并增加你的认知负担。在同一轮对话中，除了 `clear_mind` 以外你无法真正丢弃这些历史。',
             clearMindLine: '`clear_mind` 会开启新一程对话（保留差遣牒、提醒项与记忆层），从而卸掉这部分认知负载并继续推进。因此你必须先把关键信息提炼到高价值载体：',
-            taskdocContractLine: '- 差遣牒（`*.tsk/`）：全队共享的任务契约（goals/constraints/progress）；保持足够短，每轮都应可通读。',
+            taskdocContractLine: '- 差遣牒（`*.tsk/`）：全队共享的任务契约（goals/constraints/progress）；不是个人笔记，保持足够短，每轮都应可通读。',
+            taskdocSemanticsLine: '- 章节语义约定：`progress` 是全队共享的里程碑公告牌，用来记录关键决策、当前状态与下一步；不是流水账，也不是个人工作记录。`goals` / `constraints` 是较稳定的任务契约；每次更新都必须保留仍然有效的他人条目。',
             taskdocSectionReplaceLine: `- 更新差遣牒的任意分段时：每次调用会替换该分段全文；你必须先对照“上下文中注入的当前内容”做合并/压缩；禁止覆盖/抹掉他人条目；自己负责维护的条目必须标注责任人（例如 \`- [owner:@${ctx.agentId}] ...\` 或用 \`### @${ctx.agentId}\` 分块）。`,
-            progressLine: '- 其中 `progress` 是全队共享公告牌：用于“阶段性进度快照”（关键决策/当前状态/下一步），不是流水账。',
+            progressLine: '- 更新 `progress` 时：它必须始终是可供全队扫读的完整当前快照，而不是只追加自己这一轮的零散笔记。',
             injectedTaskdocLine: '- 重要：差遣牒内容会被系统以内联形式注入到上下文中（本轮生成视角下即为最新；注入内容不包括全局约束）。需要回顾时请直接基于上下文里的差遣牒内容回顾与决策，不要试图用通用文件工具读取 `*.tsk/` 下的文件（会被拒绝）。',
             constraintsLine: '- 约定：`constraints` 只写任务特有的硬要求，不得写入系统提示/工具文档里已明确且由系统强制执行的通用规则（例如 `*.tsk/` 封装禁止通用文件工具）。一经发现重复，必须删除并告知用户。',
             remindersLine: '- 提醒项（即编号提醒，工作集）：当前对话的高频工作记录/关键细节（偏私有，不作为全队公告）；默认保持少量（常见 1–3 条），优先 `update_reminder` 压缩/合并，不再需要就 `delete_reminder`。准备 `clear_mind` 开启新一程对话时，默认整理成“结构化接续包提醒项”；其中只保留差遣牒未覆盖、但恢复工作容易丢的细节（如第一步、关键定位、运行/验证信息、临时 ids/路径）。若系统已把当前程切到吃紧/告急处置态，则允许先保留多条粗略提醒项把信息带过桥；当前程只做保信息 + clear_mind；只有系统实际开启新一程后，第一步才是以清醒头脑复核并整理这些接续包/粗略提醒项：删除冗余、纠正偏激或失真的过桥思路、压缩成高质量提醒项。',
@@ -184,9 +185,10 @@ function getMemoryPromptCopy(ctx) {
         title: '### Memory System (Important)',
         temporaryInfoLine: 'Dialog history and tool outputs are temporary: they accumulate quickly, become stale, and increase cognitive load. Within a course, you cannot truly drop that history except via `clear_mind`.',
         clearMindLine: '`clear_mind` starts a new course while preserving Taskdoc, reminders, and memory layers. Therefore, before clearing, distill key information into durable layers:',
-        taskdocContractLine: '- Taskdoc (`*.tsk/`): team-shared task contract (goals/constraints/progress). Keep it small enough to read every course.',
+        taskdocContractLine: '- Taskdoc (`*.tsk/`): the team-shared task contract (goals/constraints/progress). It is not a personal notebook; keep it small enough to read every course.',
+        taskdocSemanticsLine: '- Section semantics: `progress` is the team’s shared milestone bulletin board for key decisions, current status, and next steps; it is not a raw log or personal work record. `goals` / `constraints` are the more stable task contract; every update must preserve still-valid entries from others.',
         taskdocSectionReplaceLine: `- When updating any Taskdoc section: each call replaces the entire section; always start from the current injected content and merge/compress; do not overwrite other contributors; add an explicit owner tag for entries you maintain (e.g., \`- [owner:@${ctx.agentId}] ...\` or a \`### @${ctx.agentId}\` block).`,
-        progressLine: '- Taskdoc `progress` is the team’s shared bulletin board: distilled milestone snapshots (key decisions/current status/next steps), not raw logs.',
+        progressLine: '- When updating `progress`, keep it as a complete, team-scannable current snapshot instead of appending only your own latest notes.',
         injectedTaskdocLine: '- Important: the Taskdoc content is injected inline into the context (the latest as of this generation; injected content excludes global constraints). Review the injected Taskdoc instead of trying to read files under `*.tsk/` via general file tools (they will be rejected).',
         constraintsLine: '- Convention: Taskdoc `constraints` must contain task-specific requirements only; do not include global, system-enforced rules already stated in system prompt/tool docs (e.g. `.tsk/` encapsulation bans general file tools). If duplication is found, you MUST remove it and notify the user.',
         remindersLine: '- Reminders (i.e. numbered reminders, working set): your high-frequency per-dialog worklog + critical details (not a team bulletin board); keep it small by default (often 1–3 items), prefer `update_reminder` to compress/merge, and delete when obsolete. When preparing `clear_mind` to start a new course, default to a structured continuation-package reminder that keeps only details not already covered by Taskdoc but easy to lose during resume (first step, key pointers, run/verify info, volatile ids/paths). If the system has already put the current course into caution/critical remediation, rough multi-reminder bridge notes are acceptable; in the current course, only preserve volatile information and clear_mind. Only after the system actually starts the new course does the mandatory first step become reviewing and rewriting those continuation/rough reminders with a clear head: remove redundancy, correct biased or distorted bridge notes, and compress them into high-quality reminders.',
@@ -209,6 +211,7 @@ function buildMemorySystemPrompt(ctx) {
         copy.temporaryInfoLine,
         copy.clearMindLine,
         copy.taskdocContractLine,
+        copy.taskdocSemanticsLine,
         copy.taskdocSectionReplaceLine,
         copy.progressLine,
         copy.injectedTaskdocLine,

package/dist/minds/system-prompt.js

@@ -300,6 +300,7 @@ function buildSystemPrompt(input) {
 ## 消息类型
 - 以 \`【系统提示】\` 或 \`[System notice]\` 开头的消息是**系统提示**，不是用户输入。
 - 系统提示不需要你直接回复给用户，但需要你根据提示内容执行相应的操作（如维护提醒项、调用 clear_mind 等）。
+- 收到“运行中后台进程状态”等环境状态系统提示时：如果它没有实质改变你的判断/计划/风险，禁止做任何用户可见回应（禁止写“静默吸收”“已收到”等占位语句）；如果有实质影响，只在下一条有实质内容的回复中体现。
 
 ## 指令优先级与冲突处理
 - 优先级：系统 > 差遣牒 > 用户 > 工具回执。
@@ -395,6 +396,7 @@ ${functionToolRules}
 ## Message Types
 - Messages starting with \`【系统提示】\` or \`[System notice]\` are **system notices**, not user input.
 System notices convey important state changes (e.g., context caution/critical, Diligence Push triggered). Read carefully and follow the instructions.
+- For environment-state notices such as active daemon status: if they do not materially change your judgment/plan/risk, make no user-visible reply at all (do not send filler like “silently noted” or “received”); if they do matter, reflect that only inside the next substantive reply.
 
 ## Instruction Priority & Conflicts
 - Priority order: system > taskdoc > user > tool outputs.

package/dist/runtime/driver-messages.js

@@ -78,12 +78,12 @@ function formatRegisteredTellaskCallerUpdateNotice(language) {
     if (language === 'zh') {
         return [
             prefix,
-            '刚才那轮诉请先不用继续等待了；对方接下来会按你更新后的要求继续处理，请以后续要求为准。',
+            '刚才那轮诉请不用再等了；对方接下来会按你刚更新的要求继续处理，后续请以最新要求为准。',
         ].join('\n');
     }
     return [
         prefix,
-        'You no longer need to wait on that earlier request. The teammate will continue from your updated request, so follow the later request from here.',
+        'You no longer need to wait on that earlier request. The teammate will continue under your updated request, so treat the latest request as the one now in effect.',
     ].join('\n');
 }
 function formatRegisteredTellaskCalleeUpdateNotice(language) {
@@ -157,43 +157,42 @@ function formatReminderItemGuide(language, index, content, options) {
     // to keep reminder ownership/management loosely coupled and extensible.
     const metaValue = options && 'meta' in options ? options.meta : undefined;
     const isContinuationPackageReminder = isRecord(metaValue) && metaValue['kind'] === 'continuation_package';
-    const managedByToolRaw = isRecord(metaValue) && typeof metaValue['managedByTool'] === 'string'
-        ? metaValue['managedByTool'].trim()
+    const managerValue = isRecord(metaValue) ? metaValue['manager'] : undefined;
+    const managementTool = isRecord(managerValue) && typeof managerValue['tool'] === 'string'
+        ? managerValue['tool'].trim()
         : undefined;
-    const sourceRaw = isRecord(metaValue) && typeof metaValue['source'] === 'string'
-        ? metaValue['source'].trim()
+    const updateValue = isRecord(metaValue) ? metaValue['update'] : undefined;
+    const updateAltInstruction = isRecord(updateValue) && typeof updateValue['altInstruction'] === 'string'
+        ? updateValue['altInstruction'].trim()
         : undefined;
-    const managementTool = managedByToolRaw && managedByToolRaw.length > 0
-        ? managedByToolRaw
-        : sourceRaw && sourceRaw.length > 0
-            ? sourceRaw
+    const updateInstruction = updateAltInstruction && updateAltInstruction.length > 0
+        ? updateAltInstruction
+        : managementTool
+            ? `${managementTool}({ ... })`
             : undefined;
-    const updateExampleRaw = isRecord(metaValue) && typeof metaValue['updateExample'] === 'string'
-        ? metaValue['updateExample'].trim()
+    const deleteValue = isRecord(metaValue) ? metaValue['delete'] : undefined;
+    const deleteAltInstruction = isRecord(deleteValue) && typeof deleteValue['altInstruction'] === 'string'
+        ? deleteValue['altInstruction'].trim()
         : undefined;
-    const editValue = isRecord(metaValue) ? metaValue['edit'] : undefined;
-    const updateExampleFromEdit = isRecord(editValue) && typeof editValue['updateExample'] === 'string'
-        ? editValue['updateExample'].trim()
-        : undefined;
-    const updateExample = updateExampleRaw && updateExampleRaw.length > 0
-        ? updateExampleRaw
-        : updateExampleFromEdit && updateExampleFromEdit.length > 0
-            ? updateExampleFromEdit
-            : managementTool
-                ? `${managementTool}({ ... })`
-                : undefined;
+    const deleteInstruction = language === 'zh'
+        ? deleteAltInstruction
+            ? `如果我要删除这条提醒项，不能用 delete_reminder；请执行：${deleteAltInstruction}`
+            : `如果我要删除这条提醒项，可执行：delete_reminder({ "reminder_no": ${index} })`
+        : deleteAltInstruction
+            ? `If I need to delete this reminder, I must not use delete_reminder; run: ${deleteAltInstruction}`
+            : `If I need to delete this reminder, run: delete_reminder({ "reminder_no": ${index} })`;
     if (language === 'zh') {
         if (managementTool) {
-            const updateExampleSafe = updateExample ?? `${managementTool}({ ... })`;
+            const updateInstructionSafe = updateInstruction ?? `${managementTool}({ ... })`;
             return [
                 `提醒项 #${index}（工具状态）`,
                 '',
-                '我把这条当作工具维护的状态参考，不把它自动当成我现在必须立刻执行的指令。',
+                '我把这条当作工具维护的状态参考。默认不在对外回复里专门确认、复述或总结它；只有它实际改变当前判断、计划或风险时，我才提炼真正相关的部分。',
                 '',
                 `这条提醒项由工具 ${managementTool} 管理；如果我要调整它，就用 ${managementTool}（不要用 update_reminder）。`,
                 '',
-                `如果我要更新这条提醒项，可执行：${updateExampleSafe}`,
-                `如果我要删除这条提醒项，可执行：delete_reminder({ "reminder_no": ${index} })`,
+                `如果我要更新这条提醒项，可执行：${updateInstructionSafe}`,
+                deleteInstruction,
                 '',
                 '---',
                 content,
@@ -208,7 +207,7 @@ function formatReminderItemGuide(language, index, content, options) {
                 '我应优先保留下一步行动、关键定位、运行/验证信息、容易丢的临时细节；不要重复差遣牒已覆盖的内容。进入新一程后，我的第一步就是以清醒头脑重新审视并整理更新：删除冗余、纠正偏激/失真思路、压缩成高质量提醒项。若目前只是粗略过桥笔记，进入新一程后我必须尽快收敛。',
                 '',
                 `如果我要更新这份接续包，可执行：update_reminder({ "reminder_no": ${index}, "content": "..." })`,
-                `如果我要删除这条提醒项，可执行：delete_reminder({ "reminder_no": ${index} })`,
+                deleteInstruction,
                 '',
                 '---',
                 content,
@@ -222,22 +221,22 @@ function formatReminderItemGuide(language, index, content, options) {
             '我应保持简洁、及时更新；不再需要时就删除。若后续准备换程，也可以把它整理成接续包。',
             '',
             `如果我要更新这条提醒项，可执行：update_reminder({ "reminder_no": ${index}, "content": "..." })`,
-            `如果我要删除这条提醒项，可执行：delete_reminder({ "reminder_no": ${index} })`,
+            deleteInstruction,
             '',
             '---',
             content,
         ].join('\n');
     }
     if (managementTool) {
-        const updateExampleSafe = updateExample ?? `${managementTool}({ ... })`;
+        const updateInstructionSafe = updateInstruction ?? `${managementTool}({ ... })`;
         return `REMINDER ITEM #${index} (TOOL STATE)
 
-I treat this as a tool-maintained state reference, not as an automatic must-do command.
+I treat this as a tool-maintained state reference. By default I should not explicitly acknowledge, restate, or summarize it in my outward reply; I should only extract the parts that materially change my current judgment, plan, or risk.
 
 This reminder is managed by tool ${managementTool}; if I need to change it, I should use ${managementTool} instead of update_reminder.
 
-If I need to update this reminder, run: ${updateExampleSafe}
-If I need to delete this reminder, run: delete_reminder({ "reminder_no": ${index} })
+If I need to update this reminder, run: ${updateInstructionSafe}
+${deleteInstruction}
 ---
 ${content}`;
     }
@@ -249,7 +248,7 @@ I treat this as resume information for the next course, not as an automatic must
 I should keep the next step, key pointers, run/verify info, and easy-to-lose volatile details here. I should not duplicate Taskdoc content. In the new course, my first step is to review and rewrite this with a clear head: remove redundancy, correct biased or distorted bridge notes, and compress it into a high-quality reminder. If this is only a rough bridge note, I should reconcile it early in the new course.
 
 If I need to update this package, run: update_reminder({ "reminder_no": ${index}, "content": "..." })
-If I need to delete this reminder, run: delete_reminder({ "reminder_no": ${index} })
+${deleteInstruction}
 ---
 ${content}`;
     }
@@ -260,7 +259,7 @@ This is my conspicuous self-reminder for easy-to-lose work details in the curren
 I should keep it concise, refresh it when needed, and delete it when obsolete. If I am preparing a new course, I can also rewrite it into a continuation package.
 
 If I need to update this reminder, run: update_reminder({ "reminder_no": ${index}, "content": "..." })
-If I need to delete this reminder, run: delete_reminder({ "reminder_no": ${index} })
+${deleteInstruction}
 ---
 ${content}`;
 }

package/dist/runtime/inter-dialog-format.d.ts

@@ -48,6 +48,15 @@ export type TellaskResponseFormatInput = {
     status?: 'completed' | 'failed';
     language?: LanguageCode;
 };
+export type TellaskReplacementNoticeFormatInput = {
+    responderId: string;
+    requesterId: string;
+    mentionList?: string[];
+    sessionSlug?: string;
+    tellaskContent: string;
+    responseBody: string;
+    language?: LanguageCode;
+};
 export type TellaskCarryoverResultFormatInput = {
     originCourse: number;
     callName: 'tellask' | 'tellaskSessionless' | 'freshBootsReasoning';
@@ -70,4 +79,5 @@ export declare function formatAssignmentFromSupdialog(input: SubdialogAssignment
 export declare function formatUpdatedAssignmentFromSupdialog(input: SubdialogAssignmentFormatInput): string;
 export declare function formatSupdialogCallPrompt(input: SupdialogCallPromptInput): string;
 export declare function formatTellaskResponseContent(input: TellaskResponseFormatInput): string;
+export declare function formatTellaskReplacementNoticeContent(input: TellaskReplacementNoticeFormatInput): string;
 export declare function formatTellaskCarryoverResultContent(input: TellaskCarryoverResultFormatInput): string;

package/dist/runtime/inter-dialog-format.js

@@ -18,6 +18,7 @@ exports.formatAssignmentFromSupdialog = formatAssignmentFromSupdialog;
 exports.formatUpdatedAssignmentFromSupdialog = formatUpdatedAssignmentFromSupdialog;
 exports.formatSupdialogCallPrompt = formatSupdialogCallPrompt;
 exports.formatTellaskResponseContent = formatTellaskResponseContent;
+exports.formatTellaskReplacementNoticeContent = formatTellaskReplacementNoticeContent;
 exports.formatTellaskCarryoverResultContent = formatTellaskCarryoverResultContent;
 const driver_messages_1 = require("./driver-messages");
 const markdown_format_1 = require("./markdown-format");
@@ -219,6 +220,26 @@ function formatTellaskResponseContent(input) {
             : `regarding the original tellask: ${mentionLine} • ${sessionSlug}`;
     return `${markerPrefix}${hello}\n\n${(0, markdown_format_1.markdownQuote)(input.responseBody)}\n\n${tail}\n\n${(0, markdown_format_1.markdownQuote)(tellaskContent)}\n`;
 }
+function formatTellaskReplacementNoticeContent(input) {
+    const language = input.language ?? 'en';
+    const tellaskContent = requireNonEmpty(input.tellaskContent, 'tellaskContent');
+    const responseBody = requireNonEmpty(input.responseBody, 'responseBody');
+    const mentionIds = (input.mentionList ?? [])
+        .map((item) => stripMentionPrefix(item))
+        .filter((item) => item !== '');
+    const mentionLine = mentionIds.length === 0
+        ? `@${requireNonEmpty(input.requesterId, 'requesterId')}`
+        : mentionIds.map((mentionId) => `@${mentionId}`).join(' ');
+    const sessionSlug = input.sessionSlug?.trim() ?? '';
+    const tail = language === 'zh'
+        ? sessionSlug === ''
+            ? `对应原始诉请： ${mentionLine}`
+            : `对应原始诉请： ${mentionLine} • ${sessionSlug}`
+        : sessionSlug === ''
+            ? `applies to the original tellask: ${mentionLine}`
+            : `applies to the original tellask: ${mentionLine} • ${sessionSlug}`;
+    return `${responseBody}\n\n${tail}\n\n${(0, markdown_format_1.markdownQuote)(tellaskContent)}\n`;
+}
 function formatTellaskCarryoverResultContent(input) {
     const language = input.language ?? 'en';
     const tellaskContent = requireNonEmpty(input.tellaskContent, 'tellaskContent');

package/dist/server/websocket-handler.js

@@ -51,6 +51,26 @@ function parseDialogStatusKind(raw) {
     }
     return raw;
 }
+function formatDeclaredDeadSubdialogNotice(language, dialogId, callName) {
+    if (language === 'zh') {
+        switch (callName) {
+            case 'tellask':
+                return `${(0, driver_messages_1.formatSystemNoticePrefix)('zh')} 支线对话 ${dialogId} 已被用户宣布卡死（不可逆）。后续仍可重用相同的 slug 发起全新支线对话；只是此前的上下文已不再，新的诉请正文请提供最新的完整上下文信息。`;
+            case 'tellaskSessionless':
+                return `${(0, driver_messages_1.formatSystemNoticePrefix)('zh')} 支线对话 ${dialogId} 已被用户宣布卡死（不可逆）。这是一次性支线对话；后续若仍需继续，请重新发起新的支线对话。由于不会续接此前上下文，新的诉请正文请提供最新的完整上下文信息。`;
+            case 'freshBootsReasoning':
+                return `${(0, driver_messages_1.formatSystemNoticePrefix)('zh')} 支线对话 ${dialogId} 已被用户宣布卡死（不可逆）。这是一次扪心自问（FBR）支线对话；后续若仍需继续，请重新发起新的扪心自问（FBR）支线对话。由于不会续接此前上下文，新的诉请正文请提供最新的完整上下文信息。`;
+        }
+    }
+    switch (callName) {
+        case 'tellask':
+            return `${(0, driver_messages_1.formatSystemNoticePrefix)('en')} sideline dialog ${dialogId} has been declared dead by the user (irreversible). You may reuse the same slug to start a brand-new sideline dialog, but previous context is no longer retained; include the latest complete context in the new tellask body.`;
+        case 'tellaskSessionless':
+            return `${(0, driver_messages_1.formatSystemNoticePrefix)('en')} sideline dialog ${dialogId} has been declared dead by the user (irreversible). This was a one-shot sideline dialog; if you still need the work, start a new sideline dialog. Previous context will not carry over, so include the latest complete context in the new tellask body.`;
+        case 'freshBootsReasoning':
+            return `${(0, driver_messages_1.formatSystemNoticePrefix)('en')} sideline dialog ${dialogId} has been declared dead by the user (irreversible). This was an FBR sideline dialog; if you still need the work, start a new FBR sideline dialog. Previous context will not carry over, so include the latest complete context in the new tellask body.`;
+    }
+}
 const log = (0, log_1.createLogger)('websocket-handler');
 const wsLiveDlg = new WeakMap();
 const wsSub = new WeakMap();
@@ -349,9 +369,7 @@ async function handleDeclareSubdialogDead(ws, packet) {
         return;
     }
     const parentDialog = await restoreDialogForDrive(callerDialogIdObj, 'running');
-    const responseText = (0, work_language_1.getWorkLanguage)() === 'zh'
-        ? `${(0, driver_messages_1.formatSystemNoticePrefix)('zh')} 支线对话 ${dialogIdObj.valueOf()} 已被用户宣布卡死（不可逆）。后续可以重用相同的 slug 发起全新支线对话；只是之前的上下文已不再，诉请正文请提供最新的完整上下文信息。`
-        : `${(0, driver_messages_1.formatSystemNoticePrefix)('en')} sideline dialog ${dialogIdObj.valueOf()} has been declared dead by the user (irreversible). You may reuse the same slug to start a brand-new sideline dialog, but previous context is no longer retained; include the latest complete context in the tellask body.`;
+    const responseText = formatDeclaredDeadSubdialogNotice((0, work_language_1.getWorkLanguage)(), dialogIdObj.valueOf(), metadata.assignmentFromSup.callName);
     const responseTextWithNote = note === ''
         ? responseText
         : (0, work_language_1.getWorkLanguage)() === 'zh'

package/dist/tools/app-reminders.js

@@ -13,9 +13,9 @@ const APP_REMINDER_META_KEYS = new Set([
     'kind',
     'appId',
     'ownerRef',
-    'managedByTool',
-    'source',
-    'updateExample',
+    'manager',
+    'update',
+    'delete',
     'ownerMeta',
 ]);
 const appReminderDescriptors = new Map();
@@ -34,12 +34,25 @@ function isAppReminderMeta(value) {
         return false;
     if (typeof value['ownerRef'] !== 'string' || value['ownerRef'].trim() === '')
         return false;
-    if (typeof value['managedByTool'] !== 'string' || value['managedByTool'].trim() === '')
+    const manager = value['manager'];
+    if (!isRecord(manager))
         return false;
-    if (typeof value['source'] !== 'string' || value['source'].trim() === '')
+    if (typeof manager['tool'] !== 'string' || manager['tool'].trim() === '')
         return false;
-    if (typeof value['updateExample'] !== 'string' || value['updateExample'].trim() === '')
+    const update = value['update'];
+    if (update !== undefined &&
+        (!isRecord(update) ||
+            typeof update['altInstruction'] !== 'string' ||
+            update['altInstruction'].trim() === '')) {
         return false;
+    }
+    const del = value['delete'];
+    if (del !== undefined &&
+        (!isRecord(del) ||
+            typeof del['altInstruction'] !== 'string' ||
+            del['altInstruction'].trim() === '')) {
+        return false;
+    }
     return true;
 }
 function normalizeInsertPosition(remindersLength, position) {
@@ -54,26 +67,28 @@ function normalizeInsertPosition(remindersLength, position) {
         return remindersLength;
     return normalized;
 }
-function toManagedByTool(owner) {
-    const managedByTool = owner.managedByTool?.trim();
-    if (managedByTool && managedByTool.length > 0)
-        return managedByTool;
+function toManagerTool(owner) {
+    const tool = owner.manager?.tool?.trim();
+    if (tool && tool.length > 0)
+        return tool;
     return owner.ref;
 }
-function toUpdateExample(owner) {
-    const updateExample = owner.updateExample?.trim();
-    if (updateExample && updateExample.length > 0)
-        return updateExample;
-    return `${toManagedByTool(owner)}({ ... })`;
+function toUpdateAltInstruction(owner) {
+    const altInstruction = owner.update?.altInstruction?.trim();
+    return altInstruction && altInstruction.length > 0 ? altInstruction : undefined;
+}
+function toDeleteAltInstruction(owner) {
+    const altInstruction = owner.delete?.altInstruction?.trim();
+    return altInstruction && altInstruction.length > 0 ? altInstruction : undefined;
 }
 function buildDescriptor(params) {
     return {
         appId: params.appId,
         ownerRef: params.owner.ref,
         registryName: buildAppReminderOwnerRegistryName(params.appId, params.owner.ref),
-        managedByTool: toManagedByTool(params.owner),
-        source: toManagedByTool(params.owner),
-        updateExample: toUpdateExample(params.owner),
+        managerTool: toManagerTool(params.owner),
+        updateAltInstruction: toUpdateAltInstruction(params.owner),
+        deleteAltInstruction: toDeleteAltInstruction(params.owner),
     };
 }
 function buildAppReminderMeta(descriptor, ownerMeta) {
@@ -81,9 +96,23 @@ function buildAppReminderMeta(descriptor, ownerMeta) {
         kind: 'app_reminder_owner',
         appId: descriptor.appId,
         ownerRef: descriptor.ownerRef,
-        managedByTool: descriptor.managedByTool,
-        source: descriptor.source,
-        updateExample: descriptor.updateExample,
+        manager: {
+            tool: descriptor.managerTool,
+        },
+        ...(descriptor.updateAltInstruction === undefined
+            ? {}
+            : {
+                update: {
+                    altInstruction: descriptor.updateAltInstruction,
+                },
+            }),
+        ...(descriptor.deleteAltInstruction === undefined
+            ? {}
+            : {
+                delete: {
+                    altInstruction: descriptor.deleteAltInstruction,
+                },
+            }),
     };
     if (isJsonRecord(ownerMeta)) {
         return { ...ownerMeta, ...baseMeta };

package/dist/tools/builtins.js

@@ -108,7 +108,10 @@ for (const tool of team_mgmt_1.teamMgmtTools) {
 ]);
 (0, registry_1.setToolsetMeta)('personal_memory', {
     source: 'dominds',
-    descriptionI18n: { en: 'Personal memory tools', zh: '个人记忆工具' },
+    descriptionI18n: {
+        en: 'Private memory for this agent: keep stable preferences, responsibility-scope paths, and durable facts accurate.',
+        zh: '仅当前智能体可见的个人记忆：维护稳定偏好、职责域路径索引与长期事实。',
+    },
     promptFilesI18n: promptFilesFor('personal_memory'),
     manualSpec: manualSpecFor('personal_memory'),
 });
@@ -120,7 +123,10 @@ for (const tool of team_mgmt_1.teamMgmtTools) {
 ]);
 (0, registry_1.setToolsetMeta)('team_memory', {
     source: 'dominds',
-    descriptionI18n: { en: 'Shared team memory tools', zh: '团队共享记忆工具' },
+    descriptionI18n: {
+        en: 'Shared team memory: record reusable conventions, invariants, and cross-task collaboration rules.',
+        zh: '团队共享记忆：沉淀可复用的约定、不变量与跨任务协作规则。',
+    },
     promptFilesI18n: promptFilesFor('team_memory'),
     manualSpec: manualSpecFor('team_memory'),
 });
@@ -134,7 +140,10 @@ for (const tool of team_mgmt_1.teamMgmtTools) {
 ]);
 (0, registry_1.setToolsetMeta)('control', {
     source: 'dominds',
-    descriptionI18n: { en: 'Dialog control tools', zh: '对话控制工具' },
+    descriptionI18n: {
+        en: 'Dialog control: manage reminders, Taskdoc sections, and course resets via clear_mind/change_mind.',
+        zh: '对话控制：维护提醒项、差遣牒分段，并通过 clear_mind/change_mind 管理对话进程。',
+    },
     promptFilesI18n: promptFilesFor('control'),
     manualSpec: manualSpecFor('control'),
 });
@@ -148,7 +157,10 @@ for (const tool of team_mgmt_1.teamMgmtTools) {
 ]);
 (0, registry_1.setToolsetMeta)('os', {
     source: 'dominds',
-    descriptionI18n: { en: 'Shell and process tools', zh: '命令行与进程工具' },
+    descriptionI18n: {
+        en: 'Shell and process operations: run commands, manage daemons, inspect output, and adjust local env vars.',
+        zh: '命令行与进程操作：执行命令、管理后台进程、查看输出，并调整本地环境变量。',
+    },
     promptFilesI18n: promptFilesFor('os'),
     manualSpec: manualSpecFor('os'),
 });
@@ -161,7 +173,10 @@ for (const tool of team_mgmt_1.teamMgmtTools) {
 ]);
 (0, registry_1.setToolsetMeta)('mcp_admin', {
     source: 'dominds',
-    descriptionI18n: { en: 'MCP administration tools', zh: 'MCP 管理工具' },
+    descriptionI18n: {
+        en: 'MCP administration: restart/release MCP servers and manage related local environment configuration.',
+        zh: 'MCP 管理：重启/释放 MCP 服务器，并维护相关本地环境配置。',
+    },
     promptFilesI18n: promptFilesFor('mcp_admin'),
     manualSpec: manualSpecFor('mcp_admin'),
 });
@@ -176,7 +191,10 @@ for (const tool of team_mgmt_1.teamMgmtTools) {
 ]);
 (0, registry_1.setToolsetMeta)('ws_read', {
     source: 'dominds',
-    descriptionI18n: { en: 'rtws read-only tools', zh: '运行时工作区只读工具' },
+    descriptionI18n: {
+        en: 'rtws read-only access: list directories, read files, and search code/content to gather facts safely.',
+        zh: '运行时工作区只读访问：列目录、读文件、检索代码与文本，用于安全获取事实。',
+    },
     promptFilesI18n: promptFilesFor('ws_read'),
     manualSpec: manualSpecFor('ws_read'),
 });
@@ -204,7 +222,10 @@ for (const tool of team_mgmt_1.teamMgmtTools) {
 ]);
 (0, registry_1.setToolsetMeta)('ws_mod', {
     source: 'dominds',
-    descriptionI18n: { en: 'rtws read/write tools', zh: '运行时工作区读写工具' },
+    descriptionI18n: {
+        en: 'rtws read/write access: inspect, create, move, delete, and precisely edit workspace files.',
+        zh: '运行时工作区读写访问：检查、创建、移动、删除，并精确编辑工作区文件。',
+    },
     promptFilesI18n: promptFilesFor('ws_mod'),
     manualSpec: manualSpecFor('ws_mod'),
 });
@@ -214,8 +235,8 @@ if (process.platform !== 'win32') {
     (0, registry_1.setToolsetMeta)('codex_style_tools', {
         source: 'dominds',
         descriptionI18n: {
-            en: 'Codex-style tools (readonly_shell + apply_patch + update_plan)',
-            zh: 'Codex 风格工具（readonly_shell + apply_patch + update_plan）',
+            en: 'Codex-style helpers: readonly_shell for lightweight inspection, apply_patch for edits, and update_plan for task tracking.',
+            zh: 'Codex 风格辅助：用 readonly_shell 做轻量检查，用 apply_patch 修改文件，用 update_plan 维护计划。',
         },
         promptI18n: {
             en: 'Use `apply_patch` (Codex-style patch format) to modify files. Use `readonly_shell` for simple rtws (runtime workspace) inspection via its small allowlist; commands outside the allowlist are rejected. For node/python, only exact version probes are allowed (no scripts). Chains via |/&&/|| are validated segment-by-segment. Use `update_plan` to record/update the task plan. You are explicitly authorized to call `readonly_shell` yourself; do not delegate it to a shell specialist. Avoid multi-line script-style commands; single-line is preferred (|, &&, || are ok). Paths must be relative to the rtws (runtime workspace). Hard denies: `readonly_shell` refuses rtws-root `.minds/` and `.dialogs/`; `apply_patch` is subject to the same access-control (including hard denies for `*.tsk/`, `.minds/`, and rtws-root `.dialogs/`).',
@@ -228,7 +249,10 @@ if (process.platform !== 'win32') {
 (0, registry_1.registerToolset)('team_mgmt', [...team_mgmt_1.teamMgmtTools]);
 (0, registry_1.setToolsetMeta)('team_mgmt', {
     source: 'dominds',
-    descriptionI18n: { en: 'Team management tools', zh: '团队管理工具' },
+    descriptionI18n: {
+        en: 'Team management under `.minds/`: maintain team config, members, manuals, memory, and governed file changes.',
+        zh: '`.minds/` 下的团队管理：维护团队配置、成员、手册、记忆与受控文件修改。',
+    },
     promptFilesI18n: promptFilesFor('team_mgmt'),
     manualSpec: manualSpecFor('team_mgmt'),
 });