From 22cc822bed5c873f494cd6334f54b84954e281a2 Mon Sep 17 00:00:00 2001 From: yi-ge Date: Tue, 13 Oct 2026 20:24:27 +0800 Subject: [PATCH] =?UTF-8?q?docs:=20=E5=90=8C=E6=AD=A5=E5=85=A8=E5=B1=80?= =?UTF-8?q?=E6=8F=90=E7=A4=BA=E8=AF=8D=20Markdown=20=E5=85=A8=E6=96=87?= =?UTF-8?q?=E5=8A=A0=E8=BD=BD=E8=AF=B4=E6=98=8E?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../06-agents/12-prompt-center.md | 22 +++++++++++++------ docs/04-concepts/03-agentfs.md | 2 +- .../06-agents/12-prompt-center.md | 22 +++++++++++++------ .../current/04-concepts/03-agentfs.md | 2 +- 4 files changed, 32 insertions(+), 16 deletions(-) diff --git a/docs/02-user-guide/06-agents/12-prompt-center.md b/docs/02-user-guide/06-agents/12-prompt-center.md index 2cdfb4a..0216081 100644 --- a/docs/02-user-guide/06-agents/12-prompt-center.md +++ b/docs/02-user-guide/06-agents/12-prompt-center.md @@ -10,6 +10,10 @@ keywords: [提示词中心, 系统提示词, 全局提示词, 团队提示词, P `v10.0.179+`（长期指令 `instructions.md`）。 ::: +:::info 全局提示词编辑方式 +全局提示词的普通 Markdown 全文加载暂未在正式版开放；当前正式版若显示 L0/L1/L2 表单，仍按该版本的分层规则生效。升级到支持全文加载的版本后，旧全局文件中的 L2 正文也会常驻生效，请先检查内容。 +::: + 提示词中心把分散在用户、团队和智能体目录中的提示词来源放到同一个界面中管理。打开 **资源管理器 → 提示词中心**，左侧选择作用范围，右侧编辑内容或预览组合结果。 ## 三个作用范围 @@ -22,9 +26,13 @@ keywords: [提示词中心, 系统提示词, 全局提示词, 团队提示词, P 运行时按 **全局 → 团队 → Agent 人格 → Agent 行为准则 → Agent 长期指令** 组合可编辑片段，再加入团队身份、工具、Skill、记忆和会话等运行时上下文。团队提示词不能伪造或改写真实团队身份；成员、组长和同伴关系来自团队配置。项目规则应保留在项目工作目录的 `AGENTS.md` 等文件中。 -## 分层文件的 L0 / L1 / L2 +## 全局提示词使用普通 Markdown + +全局提示词只提供一个完整 Markdown 编辑区，保存到当前用户的 `prompt.md`。非空正文会进入该用户所有智能体的系统提示词；`## L0`、`## L1`、`## L2` 等标题只作为普通标题，不改变加载范围。仅供特定任务参考的长材料适合放在独立技能或参考资料中。 + +## 团队与 Agent 分层文件的 L0 / L1 / L2 -全局和团队编辑器支持表单与 Markdown 源码两种模式；人格和原则也支持分层 Markdown。分层时必须使用严格的二级标题： +团队编辑器支持表单与 Markdown 源码两种模式；人格和原则也支持分层 Markdown。分层时必须使用严格的二级标题： ```md ## L0 @@ -43,7 +51,7 @@ keywords: [提示词中心, 系统提示词, 全局提示词, 团队提示词, P | **L1** | 默认注入系统提示词 | | **L2** | 默认不注入，保留为按需读取的详情 | -L0/L1/L2 表示加载粒度，不是权限或安全等级。分层文件没有可识别的层级标题时，会整份作为一个提示词片段加载；全局和团队编辑器可能提示格式问题。 +L0/L1/L2 表示加载粒度，不是权限或安全等级。分层文件没有可识别的层级标题时，会整份作为一个提示词片段加载；团队编辑器可能提示格式问题。 ## 长期指令使用普通 Markdown @@ -54,7 +62,7 @@ L0/L1/L2 表示加载粒度，不是权限或安全等级。分层文件没有 ## 编辑和预览 1. 选择“全局提示词”、某个团队，或展开某个智能体 -2. 分层文件按其编辑器填写 L0/L1/L2 或 Markdown；Agent 的“长期指令 instructions”直接编辑完整 Markdown +2. 全局提示词直接编辑完整 Markdown；团队分层文件可填写 L0/L1/L2 或编辑源码；Agent 文件按对应编辑器修改 3. 保存内容 4. 对 Agent 选择“提示词预览”，按全局、团队、Agent 和系统分组检查组合结果 @@ -64,7 +72,7 @@ L0/L1/L2 表示加载粒度，不是权限或安全等级。分层文件没有 - 全局只写用户对所有智能体的通用要求，团队写协作约定；个人和项目内容不要复制到可共享的 Agent 本体 - 人格写角色和表达方式，原则写边界和优先级，长期指令写稳定职责、默认策略和交付标准 -- 分层文件把每次需要的内容放在 L0/L1；长期指令全文加载，详细方法转到技能和参考资料 +- 团队、人格与原则的分层文件把每次需要的内容放在 L0/L1；全局提示词与长期指令全文加载，详细方法转到技能和参考资料 - 不要依靠提示词替代工具权限和审批策略 ## 实战示例 @@ -84,11 +92,11 @@ L0/L1/L2 表示加载粒度，不是权限或安全等级。分层文件没有 | 问题 | 原因与解决 | |------|-----------| -| 分层文件保存后关键内容没生效 | L2 默认不注入；确认每次需要的规则在 L0/L1 中 | +| 团队或 Agent 分层文件保存后关键内容没生效 | L2 默认不注入；确认每次需要的规则在 L0/L1 中 | | 长期指令保存后当前任务仍使用旧内容 | 当前执行轮次保留快照，在新的输入、显式继续或恢复接纳时读取新内容 | | 预览中看到重复内容 | 同一要求写到了多个作用范围；保留在真正归属的位置 | | 团队提示词不生效 | 确认智能体已加入该团队；多团队时检查对应来源 | -| 分层源码提示格式问题 | 检查标题是否严格为 `## L0`、`## L1`、`## L2`；长期指令无需这些标题 | +| 团队分层源码提示格式问题 | 检查标题是否严格为 `## L0`、`## L1`、`## L2`；全局提示词与长期指令无需这些标题 | | 找不到长期指令入口 | 在 **设置 → 关于** 中检查版本，升级至 `v10.0.179+` 后重试 | | 长期指令超出预算 | 精简正文，把详细步骤移到技能；它不会通过截断或 L2 隐藏来继续执行 | diff --git a/docs/04-concepts/03-agentfs.md b/docs/04-concepts/03-agentfs.md index e0561ab..df40195 100644 --- a/docs/04-concepts/03-agentfs.md +++ b/docs/04-concepts/03-agentfs.md @@ -121,7 +121,7 @@ AgentFS 的设计灵感来源于 Linux 操作系统的文件结构： ## 分层加载 -人格、原则及全局、团队分层提示词可以用 L0/L1/L2 控制加载粒度，技能等内容也支持按需发现。不是所有 AgentFS 文件都按三级加载： +人格、原则及团队分层提示词可以用 L0/L1/L2 控制加载粒度，技能等内容也支持按需发现。用户全局提示词与 Agent 长期指令使用普通 Markdown 全文加载，不按三级解析。不是所有 AgentFS 文件都按三级加载： | 层级 | 内容 | 消耗 | 用途 | |------|------|------|------| diff --git a/i18n/en/docusaurus-plugin-content-docs/current/02-user-guide/06-agents/12-prompt-center.md b/i18n/en/docusaurus-plugin-content-docs/current/02-user-guide/06-agents/12-prompt-center.md index e10c461..89d88bc 100644 --- a/i18n/en/docusaurus-plugin-content-docs/current/02-user-guide/06-agents/12-prompt-center.md +++ b/i18n/en/docusaurus-plugin-content-docs/current/02-user-guide/06-agents/12-prompt-center.md @@ -10,6 +10,10 @@ keywords: [Prompt Center, system prompt, global prompt, team prompt, Persona, Pr `v10.0.179+` (standing instructions in `instructions.md`). ::: +:::info Global prompt editor +Plain Markdown loading for the global prompt is not yet available in the released app. If your installed version still shows an L0/L1/L2 form, its layered behavior still applies. Before upgrading to a version with full-body loading, review any old global L2 content: it will become active on every run. +::: + Prompt Center brings prompt sources stored across user, team, and agent directories into one management surface. Open **Explorer → Prompt Center**, select a scope on the left, and edit or preview it on the right. ## Three Scopes @@ -22,9 +26,13 @@ Prompt Center brings prompt sources stored across user, team, and agent director Editable sections are combined in this order: **global → team → agent persona → agent principles → agent standing instructions**. Runtime context such as team identity, tools, skills, memory, and conversation is then added. A team prompt cannot forge or replace real team identity; membership and relationships come from team configuration. Keep project rules in files such as `AGENTS.md` in the project working directory. -## L0 / L1 / L2 in Layered Files +## Global Prompts Use Ordinary Markdown + +The global prompt has one editor for the complete Markdown file, saved as the current user's `prompt.md`. Its nonempty body enters the system prompt of all that user's agents. Headings such as `## L0`, `## L1`, and `## L2` are ordinary headings and do not control loading. Keep long material needed only for particular tasks in separate skills or reference files. + +## L0 / L1 / L2 in Team and Agent Files -Global and team editors support form and Markdown source modes. Persona and principles also support layered Markdown. Layered content must use exact second-level headings: +The team editor supports form and Markdown source modes. Persona and principles also support layered Markdown. Layered content must use exact second-level headings: ```md ## L0 @@ -43,7 +51,7 @@ Longer background, examples, and exceptions | **L1** | Injected by default | | **L2** | Not injected by default; available for on-demand reading | -L0/L1/L2 describes loading granularity, not permission or safety levels. A layered file without recognized headings is loaded as one complete prompt section. Global and team editors may report a format warning. +L0/L1/L2 describes loading granularity, not permission or safety levels. A layered file without recognized headings is loaded as one complete prompt section. The team editor may report a format warning. ## Standing Instructions Use Ordinary Markdown @@ -54,7 +62,7 @@ Saved changes are reread when new input, explicit continuation, or restoration i ## Edit and Preview 1. Select **Global Prompt**, a team, or expand an agent. -2. Use the relevant editor for L0/L1/L2 or Markdown in layered files. For the agent's **Standing Instructions**, edit the complete Markdown body. +2. Edit the complete Markdown body for the global prompt. Use the form or source view for layered team files, and the relevant editor for agent files. 3. Save the content. 4. Select the agent's **Prompt Preview** to inspect global, team, agent, and system groups. @@ -64,7 +72,7 @@ Preview shows sources and character counts so you can spot duplication, conflict - Put the user's general requirements in the global scope and collaboration conventions in the team scope. Keep personal and project content out of the shareable agent body. - Put role and expression in persona, boundaries and priorities in principles, and stable responsibilities, default strategies, and delivery standards in standing instructions. -- In layered files, keep content needed on every run in L0/L1. Standing instructions load in full; detailed methods belong in skills and reference material. +- In layered team, persona, and principles files, keep content needed on every run in L0/L1. Global prompts and standing instructions load in full; detailed methods belong in skills and reference material. - Do not use prompt text as a substitute for permissions or approval policies. ## Example @@ -84,11 +92,11 @@ A documentation engineer can keep each requirement in its own scope instead of p | Question | Cause and Resolution | |----------|----------------------| -| Key content in a layered file does not take effect | L2 is not injected by default; put always-needed rules in L0/L1. | +| Key content in a layered team or agent file does not take effect | L2 is not injected by default; put always-needed rules in L0/L1. | | The current task still uses old standing instructions after saving | The current execution turn keeps its snapshot. New content is read when new input, explicit continuation, or restoration is accepted. | | Preview shows duplicate content | The same requirement is stored in multiple scopes. Keep it in the scope that owns it. | | A team prompt does not take effect | Confirm membership and check the corresponding source when the agent belongs to multiple teams. | -| Layered source reports a format issue | Check for exact `## L0`, `## L1`, and `## L2` headings; standing instructions need none. | +| Team layered source reports a format issue | Check for exact `## L0`, `## L1`, and `## L2` headings; global prompts and standing instructions need none. | | Standing Instructions is unavailable | Check your version under **Settings → About**, update to `v10.0.179+`, and try again. | | Standing instructions exceed the budget | Shorten the body and move detailed steps into skills. Execution does not proceed by truncating it or hiding an L2 section. | diff --git a/i18n/en/docusaurus-plugin-content-docs/current/04-concepts/03-agentfs.md b/i18n/en/docusaurus-plugin-content-docs/current/04-concepts/03-agentfs.md index 8588438..f6d65cc 100644 --- a/i18n/en/docusaurus-plugin-content-docs/current/04-concepts/03-agentfs.md +++ b/i18n/en/docusaurus-plugin-content-docs/current/04-concepts/03-agentfs.md @@ -121,7 +121,7 @@ Made a mistake? `git revert` and it's back. Not satisfied with recent "evolution ## Tiered Loading -Persona, principles, and layered global or team prompts can use L0/L1/L2 to control loading granularity. Skills also support on-demand discovery. Not every AgentFS file uses these tiers: +Persona, principles, and layered team prompts can use L0/L1/L2 to control loading granularity. Skills also support on-demand discovery. User global prompts and agent standing instructions use ordinary Markdown loaded in full. Not every AgentFS file uses these tiers: | Tier | Content | Consumption | Purpose | |------|---------|-------------|---------|