diff --git a/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.i18n.yaml b/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.i18n.yaml
index 3d91acee93..229fffae78 100644
--- a/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.i18n.yaml
+++ b/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.i18n.yaml
@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write .agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.md
-2026-08-03-per-session-agent-presets.md: dbf2f7c3de1447382071ebcfb1d6b9abe9640210
-2026-08-03-per-session-agent-presets.zh.md: e34018f713144c6457c86d5ce4b533106a7e5372
+2026-08-03-per-session-agent-presets.md: cc40af246d018560b7bc7bea92286e0ddaf53e55
+2026-08-03-per-session-agent-presets.zh.md: 5b6113ee9f905478e59db3e982331b8faa190e3c
diff --git a/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.md b/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.md
index dbf2f7c3de..cc40af246d 100644
--- a/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.md
+++ b/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.md
@@ -23,6 +23,8 @@ Composition splits into two planes, decided by what must be shared rather than b
 
 Model routing stays out of presets. `installAgentLlmTarget` is already the per-agent seam for provider, model, and reasoning effort, and an LLM adapter mounted inside a preset would never be resolved by `agent-loop`, which lives in the host plane.
 
+The deployment ships three presets — `standard` (the full coding agent), `core-web` (a two-tool benchmark surface), and `cordis` (the standard agent plus the self-referential toolset and a composition-authoring skill).
+
 Mounting is per-session by default. Measured cost for a twelve-row composition is ~3ms and ~600KB per session, so isolation is the cheaper default than any sharing scheme, and a preset authored by a user or by an agent then has the smallest possible blast radius. A preset that genuinely owns an expensive singleton opts into sharing with Cordis's own `isolate` vocabulary: a named realm label is process-global, so two subtrees naming the same label resolve one instance.
 
 Which preset an unnamed session gets is a user setting (`agent-presets.default`) layered over the composition's own `default`, which becomes the `base`. Both layers are needed: the composition value is what a deployment ships and must keep working with no settings provider at all, and the setting is what a person changes without editing a `cordis.yml` they may not own.
@@ -43,7 +45,15 @@ Which preset an unnamed session gets is a user setting (`agent-presets.default`)
 
 **Fiber membership is object identity, not `uid`.** A `uid` is a per-registry counter, so fibers in two different roots collide on it; comparing by `uid` made one runtime's subtree answer for a service published in another. `ctx.plugin()` returns a thenable `Object.create(fiber)` wrapper that is never identical to the fiber in a parent chain, so the subtree captures its own fiber during construction.
 
-**The preset id is model-visible and must be logged.** It determines the tool set and prompt, so a resumed session has to restore the same composition; recording it is a session fact, not runtime state.
+**A preset file is an input, never a persistence target.** `EntryTree.write()` persists a tree whenever the Loader decides the config changed, and a plugin self-disposing is enough — tearing an agent down disposes its whole subtree. Inherited, that rewrites the composition it read, in practice truncating a shipped preset to `[]` the first time a session ends. The subtree overrides `write()` to do nothing.
+
+**A plugin that looks itself up in the global registry breaks inside a preset.** `ctx.tools.register()` files into the CALLING context's scope, so a plugin mounted in a preset registers for one agent and an unscoped `ctx.tools.get(name)` correctly finds nothing. `dsh-tool-skill` did exactly that and threw on every preset mount; it now compares against the definition it registered. Any plugin meant to be preset-mountable must hold its own registration rather than re-read it by name.
+
+**An entry-local `isolate` realm is invisible to the agent's own scope, not only to the host.** Only rows inside that group resolve the service. That is what makes a preset's `skills` registry belong to one agent rather than being shared — and it means a consumer left outside its provider's group silently resolves the host registry and contributes nothing.
+
+**Switching is allowed only while a session is blank.** Once a turn has run, that history was produced under the preset's tools and swapping them would strand logged tool calls, so `agentPreset.select` answers `agent-preset-locked`. A blank switch keeps the agent and the session and replaces only the subtree, because the host discards the `AgentHandle` it creates and there is no delete RPC — and keeping them is the better outcome anyway, since the session id, its workspace attachment, and its projections all stay put. The swap is unmount-then-mount (two compositions would register the same tool names into one layer), so it resolves the new preset before tearing anything down and restores the previous one when the new mount fails.
+
+**The preset id is model-visible and must be logged.** It determines the tool set and prompt, so a resumed session has to restore the same composition; recording it is a session fact, not runtime state. It rides the session header beside `cwd`, and the summary carries it so a picker shows what a session actually runs rather than the deployment's current default.
 
 ## Alternatives considered
 
diff --git a/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.zh.md b/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.zh.md
index e34018f713..5b6113ee9f 100644
--- a/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.zh.md
+++ b/.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.zh.md
@@ -23,6 +23,8 @@ Status: implemented
 
 模型路由不进 preset。`installAgentLlmTarget` 已经是 provider、model 与 reasoning effort 的按 agent 可替换点；而挂在 preset 内部的 LLM 适配器永远不会被 `agent-loop` 解析到，因为后者位于宿主平面。
 
+部署随附三个 preset —— `standard`（完整编码 agent）、`core-web`（两个工具的 benchmark 表层）与 `cordis`（标准 agent 加上自指工具集与一份组装创作 skill）。
+
 挂载默认按会话进行。实测一份十二行组装每会话约 3ms、约 600KB，因此隔离比任何共享方案都更划算；而由用户或 agent 写出的 preset 也因此拥有尽可能小的影响面。确实自带昂贵单例的 preset，可以用 Cordis 自身的 `isolate` 词汇显式选择共享：命名 realm 的 label 是进程级全局的，因此两棵子树只要写同一个 label 就解析到同一个实例。
 
 未指名 preset 的会话拿到哪一个，是一项用户设置（`agent-presets.default`），叠在组装自身的 `default` 之上——后者成为 `base`。两层都需要：组装里的值是部署交付的东西，在完全没有 settings 提供方时也必须照常工作；而设置是让人不必去改一份可能并不属于自己的 `cordis.yml` 就能调整的东西。
@@ -44,7 +46,15 @@ Status: implemented
 
 **fiber 归属判定用对象同一性，而非 `uid`。** `uid` 是按 registry 计数的序号，因此两个不同根下的 fiber 会在它上面撞号；按 `uid` 比较曾导致一个运行时的子树为另一个运行时中发布的服务背锅。`ctx.plugin()` 返回的是 thenable 的 `Object.create(fiber)` 包装对象，与父链中出现的 fiber 永远不同一，因此子树在构造时捕获自己的 fiber。
 
-**preset id 对模型可见，必须写入日志。** 它决定工具集与提示词，因此被恢复的会话必须还原同一份组装；记录它属于会话事实，而非运行时状态。
+**preset 文件是输入，绝不是持久化目标。** 只要 loader 认为配置变了，`EntryTree.write()` 就会回写整棵树，而一个插件自我 dispose 就足以触发——销毁 agent 会 dispose 它的整棵子树。若继承该行为，它会重写自己读入的那份组装，实际后果是第一次会话结束时把随附 preset 截断成 `[]`。子树因此把 `write()` 覆盖为空操作。
+
+**按自身名字回查全局注册表的插件，在 preset 里必然失效。** `ctx.tools.register()` 归档进**调用方**上下文的 scope，因此挂在 preset 里的插件只为一个 agent 注册，而不带 scope 的 `ctx.tools.get(name)` 理所当然查不到。`dsh-tool-skill` 正是这样写的，于是每次 preset 挂载都抛错；现在它与自己注册的那个定义比对。任何希望可被 preset 挂载的插件，都必须持有自己的注册对象，而不是按名字重新读取。
+
+**entry 本地 `isolate` realm 不仅对宿主不可见，对 agent 自身的 scope 同样不可见。** 只有该组内部的行能解析到该服务。这正是让 preset 的 `skills` 注册表归属单个 agent 而非共享的原因——同时也意味着：被留在提供方组之外的消费方会静默解析到宿主注册表，然后什么都不贡献。
+
+**只有空白会话才允许切换。** 一旦跑过任何轮次，那段历史就是在该 preset 的工具下产生的，替换会留下无法执行的已记录 tool call，因此 `agentPreset.select` 返回 `agent-preset-locked`。空白期的切换保留 agent 与 session，只替换子树——因为宿主丢弃了它创建的 `AgentHandle`，也没有 delete RPC；而保留它们本身就是更好的结果，会话 id、workspace 挂接与 projections 都原地不动。该替换是"先卸后装"（两份组装会把同名工具注册进同一分层），因此它在拆除任何东西之前先解析新 preset，并在新组装装载失败时恢复原来的那一份。
+
+**preset id 对模型可见，必须写入日志。** 它决定工具集与提示词，因此被恢复的会话必须还原同一份组装；记录它属于会话事实，而非运行时状态。它与 `cwd` 并列写在会话头部，并由会话摘要携带，使选择器显示的是某个会话实际运行的 preset，而非部署当前的默认值。
 
 ## 考虑过的替代方案
 
diff --git a/AGENTS.md b/AGENTS.md
index 0d27b20df0..660db23fd0 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -28,6 +28,7 @@ packages/    @deepseek-ai/dsh-<pkg> workspaces at packages/<group>/<pkg>/
   workflow/    workflow seam + worker-thread engine + workflow tool
   todo/        todo_write tool
   plan/        plan mode as logged per-agent collaboration state
+  preset/      per-session agent composition from preset cordis.yml files
   guard/       loop-hygiene plugins
   cordis/      self-referential toolset: the agent inspects/mounts plugins in its own runtime
   hooks/       Claude Code/Codex hook bridges + shared wire-protocol library
diff --git a/docs/architecture.i18n.yaml b/docs/architecture.i18n.yaml
index 0459323eea..34437b85e4 100644
--- a/docs/architecture.i18n.yaml
+++ b/docs/architecture.i18n.yaml
@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write docs/architecture.md
-architecture.md: 40c20a1c9eeabe5ecbbc6edacde81c20071b8a04
-architecture.zh.md: 6fddaa883775cf8345aba01af52575c0f0e1aaa0
+architecture.md: 3f2a879b8c3bf639b9b89b42cfac277e86d87eca
+architecture.zh.md: ed4c2662a4dfbb57fa411be6a22cd173ec56ee15
diff --git a/docs/architecture.md b/docs/architecture.md
index 40c20a1c9e..3f2a879b8c 100644
--- a/docs/architecture.md
+++ b/docs/architecture.md
@@ -164,6 +164,10 @@ Exceptions combine LLM interface/consumer, filesystem policy, web registries, an
 
 `dsh-agent-spine-demo` bundles a spine and optional goals. App packages own CLI, ACP automation, and JSON-RPC front doors ([README](../packages/examples/agent-spine-demo/README.md), [acp/](../packages/acp/README.md), [ui/](../packages/ui/README.md)). `dsh-jsonrpc-agent` boots external `cordis.yml`; the Python SDK defaults when config is absent ([Python SDK](../python/README.md)). Thin deployments use swappable backends and optional tools ([examples/](../examples/AGENTS.md), [runnable wirings](cookbook/extension-cookbook.md#runnable-wirings), [graph atlas](graph-atlas.md)).
 
+### Agent Presets
+
+A deployment may compose each session's model-facing plugin set separately. An **agent preset** is a directory holding one `agent.cordis.yml`, mounted as an `include` subtree under that agent's scope during `setup(agentCtx)`, so its tool and prompt registrations file into that agent's layer and unwind with it — no new tier in the registries. The host composition keeps what must be shared: the registries themselves, cross-session facilities, the sandbox and approval stack, the model route. `ctx.agentPresets` owns discovery and the guarded mount, rejecting a row that never activates or that publishes into the root service realm. Details: [per-session agent presets](../.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.md), [preset/](../packages/preset/README.md).
+
 ### Where New Behavior Goes
 
 New behavior attaches to a documented extension point; a loop change updates this map.
@@ -172,6 +176,7 @@ New behavior attaches to a documented extension point; a loop change updates thi
 |---|---|
 | Add a model provider | register its adapter on `ctx.llm` |
 | Add a model-facing capability | register on `ctx.tools`; schemas join prompt assembly |
+| Give one session a different capability set | compose it in an agent preset; a service row there needs an `isolate` realm |
 | Add shell execution | implement and register a `ctx.bash` backend; the local backend spawns through `ctx.subprocess` |
 | Add persistent terminal execution | register a `ctx.pty` backend plus `dsh-tool-pty` |
 | Add a human command | register on `ctx.commands`; adapters discover and dispatch without a model turn |
diff --git a/docs/architecture.zh.md b/docs/architecture.zh.md
index 6fddaa8837..ed4c2662a4 100644
--- a/docs/architecture.zh.md
+++ b/docs/architecture.zh.md
@@ -164,6 +164,10 @@ idle inject:
 
 `dsh-agent-spine-demo` 组合一套主干和可选目标。应用包负责 CLI（命令行界面）、ACP 自动化入口和 JSON-RPC 入口（[README](../packages/examples/agent-spine-demo/README.md)、[acp/](../packages/acp/README.md)、[ui/](../packages/ui/README.md)）。`dsh-jsonrpc-agent` 启动外部 `cordis.yml`；Python SDK 在配置缺失时提供默认项（[Python SDK](../python/README.md)）。轻量部署使用可替换后端和可选工具（[examples/](../examples/AGENTS.md)、[可运行接线](cookbook/extension-cookbook.md#runnable-wirings)、[图谱](graph-atlas.md)）。
 
+### Agent Preset
+
+部署可为每个会话分别组装面向模型的插件集合。**agent preset** 是一个含 `agent.cordis.yml` 的目录，在 `setup(agentCtx)` 期间作为 `include` 子树挂到该 agent 的 scope 之下，其工具与提示词注册因而归档进该 agent 的分层并随之卸载，注册表无需新增层级。宿主组装保留必须共享的部分：注册表本身、跨会话设施、沙箱与审批栈、模型路由。`ctx.agentPresets` 负责发现与把关，拒绝未激活的行和把服务发布进根 realm 的行。详见 [按会话组装 agent preset](../.agents/notes/implemented/architecture/2026-08-03-per-session-agent-presets.md)、[preset/](../packages/preset/README.md)。
+
 ### 新行为的归属位置
 
 新行为附加到已有文档记录的扩展点；循环发生变更时，本架构图随之更新。
@@ -172,6 +176,7 @@ idle inject:
 |---|---|
 | 添加模型提供方 | 在 `ctx.llm` 上注册其适配器 |
 | 添加面向模型的功能 | 在 `ctx.tools` 上注册；schema 加入提示词组装 |
+| 让某个会话拥有不同的能力集合 | 在 agent preset 中组装它；其中的 service 行需要 `isolate` realm |
 | 添加 shell 执行 | 实现并注册 `ctx.bash` 后端；本地后端通过 `ctx.subprocess` 生成进程 |
 | 添加持久化终端执行 | 注册 `ctx.pty` 后端和 `dsh-tool-pty` |
 | 添加用户命令 | 在 `ctx.commands` 上注册；适配器无需模型轮次即可发现并分派 |
diff --git a/scripts/doc-budgets.manifest.json b/scripts/doc-budgets.manifest.json
index 1ca6dac5e1..42d1d4c90b 100644
--- a/scripts/doc-budgets.manifest.json
+++ b/scripts/doc-budgets.manifest.json
@@ -1,7 +1,7 @@
 {
-  "AGENTS.md": 1775,
+  "AGENTS.md": 1900,
   "docs/AGENTS.md": 1320,
-  "docs/architecture.md": 2160,
+  "docs/architecture.md": 2400,
   "docs/cordis-primer.md": 600,
   "docs/defensive-patterns.md": 550,
   "docs/testing.md": 1150,