fix(preset): keep the token meter host-plane and name unjoined agents

Moving the agent plane behind presets left two readers on the wrong side of
the host/agent line.

`dsh-token-meter` was disabled on the host and mounted inside each preset's
`compaction` realm, but its three projection units register into the
process-wide `sessionProjections` table. A unit registered from one preset
answers for every session, so whether a `minimal` session showed a context
meter depended on whether some other session had mounted `standard` since
boot, and a process that only ever ran `minimal` showed none. The meter takes
no configuration, keys every fold by Session, and registers no tool or prompt
section, so it returns to the host composition and leaves the presets'
`isolate` map; the realm and `compact-basic` stay, because what a preset
chooses is whether its agent compacts, not whether its tokens are counted.

Nothing named an agent that joined no preset. The join is a scope-parent link,
and without it the tools, prompt-section, and skill views resolve the empty
global layer: the agent publishes, the turn runs, and the model receives
nothing. `AgentPresets` now logs one warning per such agent while a roster is
configured, and the invariant companion fails outright — at
`system-prompt/assemble` rather than at publication, because an unjoined agent
is legal until it addresses a model and `recompose` binds exactly such an
agent. The warning stays advisory: a synchronous `agent/created` throw vetoes
publication, and the ACP bridge, SDK server, and headless bundle all create an
unjoined agent today.

Three limits are recorded rather than fixed: projection key presence is not a
per-session capability signal, a superseded standing generation is never
reclaimed, and a `cordis_mount` temporary plugin belongs to the composition
rather than the session that mounted it.

Fixes #2203

packages/preset/agent-presets/README.zh.md

@@ -133,7 +133,8 @@ Indirectly, through the plugins a standing composition registers, which own ever
 ## Known Limitations and Deferred Work
 
 - **会话一旦产出内容便无法更换 preset** —— `recompose` 把**空白**会话的父作用域重链到另一个常驻挂载，且仅限空白会话：切换已运行过的组装会抽走模型已调用的工具。更改默认值只影响此后创建的会话。
-- **代际只以组装文件为键** —— stamp 检查只察觉 `agent.cordis.yml` 的变化，察觉不到旁边 skill 文件或资产的编辑；那些编辑要等组装文件本身变动或进程重启才达到新会话。已加入的会话保持其代际，进程存活期间不回收被替代的代际（上限取决于组装被编辑的频率，而非会话数）。
+- **代际只以组装文件为键** —— stamp 检查只察觉 `agent.cordis.yml` 的变化，察觉不到旁边 skill 文件或资产的编辑；那些编辑要等组装文件本身变动或进程重启才达到新会话。
+- **被替代的代际永不回收** —— 已加入的会话保持其运行所在的代际，而名单没有加入计数可以判断最后一个何时离开，因此整棵子树一直挂到进程结束。代价按代际计而非按会话计，但并非为零：`dsh-skill-local` 默认监听自己的根目录，因此每一轮「编辑后建会话」都会新增一套活的 watcher。上限取决于组装被编辑的频率——而设置页的编写流程把这件事从「每次部署」变成了「每次保存」。要回收就需要给常驻挂载加上已加入 agent 的计数；见 `ensureStanding` 处的 `TODO`。
 - **副本从不被实际挂载以校验** —— 它与来源逐字节相同，因此磁盘上已坏的来源会产出与来源同样损坏的副本；发现过程的健康检查会在下一次读取名单时把两行都标出来，而不是把失败推迟到会话启动。
 - **健康是形状检查，不是挂载** —— 发现过程只证明组装能以加载器方言解析、由具名行组成，不证明每一行的模块都能解析并激活；引用不存在的包的行仍在第一个会话处失败，并回滚该会话的创建。
 - **副本是会漂移的快照** —— 升级部署不会更新随附 preset 的副本，本层也没有表达「standard 加一处改动」的 patch 语义（那是 bundle 层 `cordis.patch.yml` 的能力）；随附集合自己也接受同样的代价——`cordis` 与 `code` 就是 `standard` 的完整副本——换来整份组装在一个文件里可读。