feat(dashboard): LLM stats answer who/what spends tokens — member scope chips and spend breakdowns

The stats section was designed single-user: four global charts, no
attribution. Request metadata rows now carry the session's source and
the frame's agent_id/depth (additive ensure_column), denormalized at log
time by the LoggingModel from the owner's pool — off the turn's hot
path, degrading to NULLs, never to a lost row.

The dashboard gains member scope chips filtering every chart, and a
breakdown row splitting the range's billed tokens by member, kind
(chat / sub-agents / cron / system agents / channels), agent, model and
provider — a sub-agent's spend is attributed to the sub-agent itself.
Rows predating the columns group under 'older data'.

docs/dashboard.md

@@ -15,17 +15,27 @@ The status reflects the **whole instance**, not one person's account: there is o
 
 ## LLM usage stats
 
-Four charts with a range switch (last hour / 24 hours / 7 days / 30 days):
+The section answers two questions: *how much is the instance being used?* and *who or what is spending the tokens?* A range switch (last hour / 24 hours / 7 days / 30 days) applies to everything on it, and — on an instance with more than one member — a row of member chips filters **all** the charts to one person; **Everyone** is the whole instance, as before.
+
+Three charts show the trend over the range:
 
 - **Requests** — how many LLM calls per minute, hour or day.
 - **Tokens** — the metered volume, split into input (split again into cached and non-cached) and output. The tooltip shows the cache-hit percentage: repeated context that was *cached* costs less and answers faster, so a high hit rate is good news, not a sign something is stuck.
 - **Avg latency** — how long a model call took on average.
-- **Models** — the top models by requests in the range.
+
+Below them, **How the spend splits** breaks the range's billed tokens down into bars:
+
+- **By member** — who consumed what (hidden while a member chip is selected: it would be a single bar).
+- **By kind** — direct chats vs **sub-agents** (specialists the assistant delegates to) vs **scheduled tasks** and the other background system agents. Data from before this breakdown existed groups under *Older data*.
+- **By agent** — which assistant or specialist consumed the most.
+- **By model** and **By provider** — where the money actually goes.
+
+A bar's tooltip shows the request count and the input/output/cached split.
 
 Three honest answers to give with a straight face:
 
-- **These numbers are everyone's, together.** The charts aggregate the whole instance; there is no per-person breakdown on this page.
-- **They record how much, when and which model — never what was said.** The content of a request lives in the requester's own encrypted space; the charts read only counters.
+- **They record how much, when, by whom and on which model — never what was said.** The content of a request lives in the requester's own encrypted space; the charts read only counters.
+- **The per-member chips are a consumption view, not a surveillance tool.** They show token counts, not conversations.
 - **Empty is normal on a new instance.** "No LLM requests in the selected range" means exactly that: nothing has run in that window.
 
 ## Pending
@@ -41,12 +51,12 @@ The same cards as the [Inbox](inbox.md) — approvals, questions and sign-in pro
 
 - **Not a monitor.** Nothing here alerts anyone; it shows the present state to whoever is looking.
 - **Not where models are configured.** Adding providers and models, and their priority order, is the admin's Models and Providers pages.
-- **Not per-person.** No page on the instance shows "who used how much" — deliberately; usage is shared, like the models.
+- **Not a message log.** The per-member and per-kind charts show volumes of tokens, never what anyone asked or was answered.
 
 ## Common questions
 
 - *"It says Degraded — should I worry?"* — it means the model checks are not all passing. Individual chats may still work on a fallback model; if it persists, the admin checks the provider (its key, its quota) on the Models/Providers pages.
-- *"Why are the bars so high at odd hours?"* — scheduled background work (system agents, cron tasks) uses the same models. The Tasks page and the system-agents page show what ran when.
+- *"Why are the bars so high at odd hours?"* — scheduled background work (system agents, cron tasks) uses the same models. The **By kind** chart shows exactly how much of the spend is background work versus direct chats; the Tasks page and the system-agents page show what ran when.
 - *"What is a token?"* — the unit LLM providers meter and bill by, roughly a word fragment. Input is what was sent (long history = more input; caching repeats cheaply), output is what was written back.
 - *"Why doesn't my child see this page?"* — their role uses the simple interface: chat, inbox and projects only.
 - *"Does the dashboard show what people asked?"* — no. Only counts, timings and model names; never content.