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per-session capability signal, a superseded standing generation is never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

apps/cli/config/agent-presets/code/agent.cordis.yml

@@ -134,17 +134,20 @@
 
 # `compact-basic` reads `toolResultPrune` through `ctx.get`, so the pruner must
 # share this realm rather than sit outside it.
+#
+# `tokenMeter` is deliberately NOT in this realm: the meter stays on the HOST
+# plane, and the rows here resolve that one instance. It takes no configuration,
+# keys every fold by Session, and owns the context-meter projection units the
+# browser reads for every session — behind a realm those units would come and go
+# with whichever presets happen to be mounted. What a preset chooses is whether
+# its agent compacts at all, which is `compact-basic` below.
 - id: compaction
   name: cordis:group
   group: true
   isolate:
-    tokenMeter: true
     compact: true
     toolResultPrune: true
   config:
-    - id: token-meter
-      name: '@deepseek-ai/dsh-token-meter'
-
     - id: compact-basic
       name: '@deepseek-ai/dsh-compact-basic'
 

apps/cli/config/agent-presets/cordis/agent.cordis.yml

@@ -115,17 +115,20 @@
 
 # `compact-basic` reads `toolResultPrune` through `ctx.get`, so the pruner must
 # share this realm rather than sit outside it.
+#
+# `tokenMeter` is deliberately NOT in this realm: the meter stays on the HOST
+# plane, and the rows here resolve that one instance. It takes no configuration,
+# keys every fold by Session, and owns the context-meter projection units the
+# browser reads for every session — behind a realm those units would come and go
+# with whichever presets happen to be mounted. What a preset chooses is whether
+# its agent compacts at all, which is `compact-basic` below.
 - id: compaction
   name: cordis:group
   group: true
   isolate:
-    tokenMeter: true
     compact: true
     toolResultPrune: true
   config:
-    - id: token-meter
-      name: '@deepseek-ai/dsh-token-meter'
-
     - id: compact-basic
       name: '@deepseek-ai/dsh-compact-basic'
 

apps/cli/config/agent-presets/minimal/agent.cordis.yml

@@ -7,7 +7,9 @@
 # from a shared default.
 #
 # The host composition is unchanged: this agent still runs inside the same
-# sandbox, approval, persistence, and model routing as any other session.
+# sandbox, approval, persistence, and model routing as any other session — and
+# the same token meter, so the browser's context meter reads this session like
+# any other. What this preset drops is auto-compaction, not the accounting.
 
 - id: persona
   name: '@deepseek-ai/dsh-persona'

apps/cli/config/agent-presets/standard/agent.cordis.yml

@@ -127,17 +127,20 @@
 
 # `compact-basic` reads `toolResultPrune` through `ctx.get`, so the pruner must
 # share this realm rather than sit outside it.
+#
+# `tokenMeter` is deliberately NOT in this realm: the meter stays on the HOST
+# plane, and the rows here resolve that one instance. It takes no configuration,
+# keys every fold by Session, and owns the context-meter projection units the
+# browser reads for every session — behind a realm those units would come and go
+# with whichever presets happen to be mounted. What a preset chooses is whether
+# its agent compacts at all, which is `compact-basic` below.
 - id: compaction
   name: cordis:group
   group: true
   isolate:
-    tokenMeter: true
     compact: true
     toolResultPrune: true
   config:
-    - id: token-meter
-      name: '@deepseek-ai/dsh-token-meter'
-
     - id: compact-basic
       name: '@deepseek-ai/dsh-compact-basic'
 

apps/cli/package.json

@@ -36,6 +36,7 @@
     "@deepseek-ai/dsh-pty-local": "workspace:^",
     "@deepseek-ai/dsh-pwsh-local": "workspace:^",
     "@deepseek-ai/dsh-pwsh-sandbox": "workspace:^",
+    "@deepseek-ai/dsh-session-projection": "workspace:^",
     "@deepseek-ai/dsh-session-reference": "workspace:^",
     "@deepseek-ai/dsh-skill": "workspace:^",
     "@deepseek-ai/dsh-skill-local": "workspace:^",

apps/cli/tests/web-agent-presets.e2e.ts

@@ -15,6 +15,9 @@ import { applyChildComposition, childSessionMeta } from '@deepseek-ai/dsh-subage
 import { CallId } from '@deepseek-ai/dsh-llm'
 import type {} from '@deepseek-ai/dsh-skill'
 import type {} from '@deepseek-ai/dsh-tools'
+// Type-only: resolves `ctx.get('sessionProjections')` and `ctx.get('tokenMeter')`.
+import type {} from '@deepseek-ai/dsh-session-projection'
+import type {} from '@deepseek-ai/dsh-token-meter'
 
 const CONFIG_DIR = fileURLToPath(new URL('../config/', import.meta.url))
 const REPO_ROOT = fileURLToPath(new URL('../../..', import.meta.url))
@@ -113,6 +116,31 @@ describe('the shipped Web composition', () => {
     expect(toolNames(ctx)).toEqual([])
   })
 
+  it('keeps the token meter and its context-meter projections on the host plane', async () => {
+    // Read before any preset in this file mounts, which is what makes this an
+    // ownership assertion rather than a mount-order coincidence: a preset-side
+    // meter sits behind an `isolate` realm and is invisible to `ctx.get`.
+    //
+    // The projection registry is process-wide rather than scope-layered, so a
+    // preset-side meter would also make the browser's context meter appear for
+    // a `minimal` session the moment some OTHER session mounted a preset that
+    // carries one, and vanish entirely in a process that only ever ran
+    // `minimal`. Host ownership is what makes the meter a per-session fact.
+    expect(ctx.get('tokenMeter')).toBeDefined()
+    const projections = ctx.get('sessionProjections')
+    if (projections === undefined) throw new Error('the Web composition must compose a projection registry')
+    const handle = await ctx.agents.create({
+      sessionId: SessionId('preset-minimal-meter'),
+      setup: agentCtx => ctx.agentPresets.mount(agentCtx, 'minimal').then(() => undefined),
+    })
+    try {
+      expect(Object.keys(projections.snapshot(handle.agent.session).values).sort())
+        .toEqual(expect.arrayContaining(['contextBreakdown', 'contextPressure', 'tokenUsage']))
+    } finally {
+      await handle.dispose()
+    }
+  })
+
   it('supplies both shipped presets, and only those, from the system root', async () => {
     const listed = await ctx.agentPresets.list()
 

apps/web/tests/seeded-history.e2e.ts

@@ -18,7 +18,6 @@ import { afterAll, beforeAll, describe, expect, it, onTestFailed } from 'vitest'
 import { createUserMessage } from '@deepseek-ai/dsh-llm'
 import type { ContentBlock, Message } from '@deepseek-ai/dsh-llm'
 import { deriveEventMessage, SessionId } from '@deepseek-ai/dsh-session'
-import type {} from '@deepseek-ai/dsh-agent-presets'
 import type { SessionEvent } from '@deepseek-ai/dsh-session'
 import type { TokenMeterService } from '@deepseek-ai/dsh-token-meter'
 import { join } from 'node:path'
@@ -195,21 +194,11 @@ describe('web e2e: seeded history renders through cold resume', () => {
     if (MODE !== 'record') {
       const raw = await readFile(SEED, 'utf8')
       expect(fixtureUserPrompts(raw), 'seed fixture must carry exactly the drive prompt').toEqual([PROMPT])
-      // The meter belongs to an agent's preset, not to the process — token
-      // accounting is per session. It is used here as a pure pricing function
-      // over fixture content, so a throwaway composition is enough to reach one.
-      const priced = await scaffold.ctx.agents.create({
-        sessionId: SessionId('seeded-history-pricing'),
-        setup: agentCtx => scaffold.ctx.agentPresets.mount(agentCtx).then(() => undefined),
-      })
-      let realizedWithCompaction: string
-      try {
-        const meter = scaffold.ctx.agentPresets.serviceFor(priced.agent, 'tokenMeter')
-        if (meter === undefined) throw new Error('seeded-history requires the composed token meter')
-        realizedWithCompaction = withCompaction(realizeSeedFixture(scaffold, raw, SEED_ID), meter)
-      } finally {
-        await priced.dispose()
-      }
+      // The meter is host-plane — it takes no configuration and keys every
+      // fold by Session — so pricing fixture content needs no agent at all.
+      const meter = scaffold.ctx.get('tokenMeter')
+      if (meter === undefined) throw new Error('seeded-history requires the host token meter')
+      const realizedWithCompaction = withCompaction(realizeSeedFixture(scaffold, raw, SEED_ID), meter)
       await seedSession(scaffold, realizedWithCompaction, SEED_ID)
     }
     browser = await chromium.launch()