Drop COMMANDCODE_DEBUG, COMMANDCODE_DEBUG_FILE, and COMMANDCODE_ZDR env vars

Sidebar (src/tui.tsx):
- Read quota()/now() signals inside the returned JSX instead of hoisting
  quota() into a const above the return. The hoisted read subscribed once
  under a non-tracking Solid owner, freezing the panel after first paint;
  the slot handler re-runs on session change, which masked the bug as
  '/session fixes it'.
- Harden refreshQuota: whole refresh (key resolve + fetch) inside
  try/catch so any throw surfaces inline instead of wedging on loading.
  Defensive api.state access (non-array provider tolerated) and
  process.env access via safeEnv; key source tracked by name, never logged.
- Refresh triggers: keep session.idle, add session.status/session.updated/
  message.updated (debounced) plus refresh on sidebar session_id change.
- Debug-gated tui-quota trace lines to the shared debug file (silent
  unless debug is on); /cc-usage hardened the same way.

Provider cost (src/model.ts, src/usage.ts):
- Forward per-request USD from finish-step (usage.raw cost/market_cost/
  gateway_cost) and provider-metadata (gateway.cost/marketCost) into
  providerMetadata.commandcode.{cost,marketCost} on the terminal finish
  part and doGenerate result; also stash into usage.raw. A reported 0 is
  preserved; only undefined means absent.

Docs: README sidebar interval corrected (3 min, not 60 s) and trigger
list updated; AGENTS.md invariants 10 (cost) and 13 (Solid gotcha).

AGENTS.md

@@ -23,14 +23,15 @@ Run from the repository root.
 
 ```powershell
 npm install        # dev deps: typescript, @types/node, @ai-sdk/provider
-npm run typecheck  # tsc --noEmit
-npm run build      # tsc -> dist/
+npm run typecheck  # tsc --noEmit (provider) + tsc -p tsconfig.tui.json (TUI plugin)
+npm run build      # tsc -> dist/ (provider only; the TUI plugin is loaded from source)
 npm run smoke      # live request against CommandCode (needs a key)
+npm run quota      # live 5-hour/weekly/monthly quota (needs a key)
 npm run sync-models # regenerate provider.commandcode.models from the live catalog
 ```
 
 There is no unit-test suite. `scripts/smoke.mjs` is the end-to-end check. `scripts/sync-models.mjs`
-is the model-catalog generator (needs `dist/` first; writes the opencode config unless `--out`/`--dry-run`).
+is the interactive model picker (needs `dist/` first; needs a TTY; writes the global opencode config).
 
 After any change under `src/`, run `npm run typecheck`, then `npm run build`, then `npm run smoke`.
 A change is not done until it typechecks and the smoke test passes.
@@ -44,8 +45,15 @@ src/transform.ts  LanguageModelV3CallOptions -> /alpha/generate envelope (return
 src/events.ts     Async iterator over the NDJSON/SSE response body.
 src/usage.ts      finish event -> LanguageModelV3Usage; finish-reason unification.
 src/redact.ts     Credential scrubbing for error surfaces.
+src/log.ts        Opt-in tracing (toggle file) to a log file.
+src/toggles.ts    Shared toggle file (~/.config/opencode/commandcode-toggles.json).
+src/quota.ts      Live 5-hour/weekly/monthly quota from the CommandCode alpha billing API.
+src/tui.tsx       opencode TUI plugin: /cc-zdr, /cc-debug, /cc-status, /cc-usage + sidebar.
+src/tui-shims.d.ts Ambient types for the opencode-provided solid-js/@opentui/solid runtime.
+tsconfig.tui.json Typechecks src/tui.tsx (+ quota/toggles) against the shim (no dependencies).
 src/constants.ts  Defaults, paths, headers, passthrough params, static config block.
 scripts/smoke.mjs Live end-to-end check.
+scripts/quota.mjs Print the live quota headlessly (reads the same config as smoke).
 scripts/sync-models.mjs Catalog -> provider.commandcode.models generator (with vision probing).
 server.py         Reference Python proxy (do not modify unless explicitly asked).
 ```
@@ -99,6 +107,7 @@ These are load-bearing. Breaking one causes silent failures in opencode.
    be omitted from the envelope; upstream rejects the whole request otherwise.
 6. **Usage is nested in V3.** Fill `inputTokens.{total,noCache,cacheRead,cacheWrite}` and
    `outputTokens.{total,text,reasoning}` — not flat `promptTokens`/`completionTokens`.
+   `inputTokens.total` is input-only (parity with `server.py`); `totalTokens` is input+output.
 7. **`supportedUrls` is `{}` and stays empty.** CommandsCode URLs are not fetched by the SDK.
 8. **Redact before surfacing errors.** Any upstream error text passed to the client must go through
    `redact()`.
@@ -107,12 +116,68 @@ These are load-bearing. Breaking one causes silent failures in opencode.
    `transform.ts`, forward each assistant `reasoning` part as `{type:"reasoning", text, signature?}`;
    do not drop it and do not fabricate empty reasoning. `server.py` predates this requirement and is
    not the guide here.
-10. **`file://` npm specs bypass install.** opencode imports `dist/index.js` directly, so the repo
+10. **Upstream cost never appears on `finish`.** Per-request USD arrives on `finish-step`
+    (`usage.raw.cost/market_cost/gateway_cost`, numbers) and `provider-metadata`
+    (`providerMetadata.gateway.cost/marketCost`, strings). `finish.totalUsage` carries tokens
+    only. Capture both into `providerMetadata: { commandcode: { cost, marketCost } }` on the
+    terminal V3 `finish` part (and `doGenerate` result), and stash into `usage.raw`. A reported
+    `0` is meaningful; only `undefined` means absent. The built-in sidebar still shows `$0.00`
+    until the fork consumes this field (anomalyco/opencode#43818); `/cc-usage` is the accurate
+    dollar source meanwhile.
+11. **`file://` npm specs bypass install.** opencode imports `dist/index.js` directly, so the repo
     must be rebuilt for opencode to see source changes.
-11. **`/models` is config-driven, not provider-driven.** opencode builds the model list from
+12. **`/models` is config-driven, not provider-driven.** opencode builds the model list from
     `provider.commandcode.models` in `opencode.json` and never asks a custom `file://` provider to
     discover models (only internal providers can register `discoverModels`). Refresh the map with
     `npm run sync-models`; do not expect a discovery hook in `src/` to populate it.
+ 13. **The TUI plugin lives outside the provider entrypoint.** `src/tui.tsx` is loaded from
+    **source** (not `dist/`) via a `tui.json` `plugin` entry that must be a **direct `file://` URL
+    to `src/tui.tsx`** (global `~/.config/opencode/tui.json` or project `.opencode/tui.json`),
+    wired by hand: `opencode plugin <path>` cannot install a bare `.tsx` (it needs a `package.json`
+    manifest next to the target), and the spec must never point at the **repo root** — a directory
+    spec is also picked up as a server plugin via `package.json` `main` → `dist/index.js` (the
+    provider, not a plugin). Do not re-add a `./tui` export for this; it is deliberately removed.
+    opencode compiles the `.tsx` at load with
+    its Bun/Solid transform and maps `solid-js` / `@opentui/solid` to its internal modules, so the
+    package stays dependency-free — do not add a build step or real imports of those packages.
+    `src/tui-shims.d.ts` is the only local stand-in for types; keep it in sync with what the file
+    actually imports. Verified on opencode 1.18.31: the module's **default export must be
+    `{ id, tui }`** (a bare `tui` named export is imported but never invoked), `id` is required,
+    and `tui` is `async (api) => {}`. Commands register via
+    `api.keymap.registerLayer({ commands })` with `{ name, run, title, desc, category,
+    namespace: "palette", slashName }`. The sidebar panel registers via
+    `api.slots.register({ order, slots: { sidebar_content } })`; `order: 90` keeps it above the
+    built-in panels (100 context, 200 mcp, 300 lsp, 400 todo, 500 files). The slot renderer
+    returns Solid JSX and is reactive; it relies on `api.theme` and `api.lifecycle.onDispose`.
+    Solid reactivity gotcha: **read `createSignal` getters inside the returned JSX, never hoist
+    `signal()` into a `const` above the `return` in the slot component.** A hoisted read subscribes
+    once under a non-tracking owner and the panel freezes after first paint (fetch succeeds, Solid
+    never re-renders). The slot handler itself re-runs on session change, which masks the bug —
+    it looks "fixed" after `/session` but stays stale on fresh launch. The built-in
+    `internal:sidebar-context` plugin reads its memos inside JSX for the same reason.
+14. **Slash commands and `Ctrl+P` share one registry.** In opencode 1.x the slash menu queries
+    the same `namespace: "palette"` commands the palette lists (both filter out `hidden: true`).
+    A slash-only entry is not expressible; registering `/cc-*` also adds them to `Ctrl+P`.
+15. **Toggles are resolved per request, never at load.** `zdr` and `debug` are read in
+    `model.ts`/`log.ts` on every call from the toggle file, so `/cc-*` changes take effect
+    without restarting opencode. Precedence for `zdr`: `providerOptions.commandcode.zdr` >
+    `x-cmd-zdr` header > toggle file. Do not reintroduce load-time consts.
+16. **Quota is live from the alpha billing API, not `server.py`.** `src/quota.ts` reads
+    `/alpha/whoami`, `/alpha/billing/credits`, `/alpha/billing/subscriptions`, and
+    `/alpha/usage/summary` (the same endpoints the `cmd` CLI `/usage` uses). `server.py` predates
+    these and is not the guide here. `resetAt` has shipped as both seconds and epoch ms — normalize
+    (>= 1e12 = ms). The monthly meter is **derived** (spend + remaining credits), so it works on any
+    plan; never hardcode a plan→cap table. The TUI resolves the key from `api.state` (provider
+    entry, then `config.provider.commandcode.options`, then `COMMANDCODE_API_KEY`), expanding
+    `{env:VAR}` itself, and must never log the key — quota errors go through `redact()`.
+17. **Retryable upstream failures can arrive inside an HTTP 200 stream.** CommandCode's gateway
+    answers `200` and then emits an SSE `error` event carrying `statusCode`/`isRetryable` (e.g.
+    `{"type":"server_error","message":"Invalid error response format: Gateway request failed",
+    "statusCode":520,"isRetryable":true}`). `fetchWithRetry` buffers only the `start`/`start-step`
+    preamble and, if the first real event is such a retryable error and attempts remain, cancels the
+    body and retries with backoff. Once any content event is seen the error is surfaced, not
+    retried (retrying mid-stream would duplicate output). Non-retryable in-stream errors (400s)
+    surface immediately. Keep the preamble set in sync if upstream adds new pre-content events.
 
 ## Change workflow
 
@@ -135,6 +200,12 @@ Automated:
 npm run typecheck; npm run build; npm run smoke
 ```
 
+Quota changes additionally need a live check (needs a key):
+
+```powershell
+npm run quota
+```
+
 Inside opencode (restart it first if config changed):
 
 ```powershell

README.md

@@ -48,7 +48,7 @@ npm run build
   "provider": {
     "commandcode": {
       "name": "Command Code",
-      "npm": "file:///C:/DevTools/pienv/ccprovider/dist/index.js",
+      "npm": "file:///C:/dev/opencode-commandcode-provider/dist/index.js",
       "options": {
         "apiKey": "{env:COMMANDCODE_API_KEY}"
       },
@@ -84,13 +84,26 @@ to it, but opencode builds `/models` and the TUI picker from this map and never
 
 opencode reads the model list from `provider.commandcode.models`; a custom provider cannot
 register dynamic discovery. `scripts/sync-models.mjs` fetches the live catalog from
-`GET /provider/v1/models` and writes the full map into your opencode config.
+`GET /provider/v1/models` and opens an interactive checkbox picker (nothing pre-checked).
+Only the selected models are vision-probed and written into your global opencode config —
+deselected ids are removed from it.
 
 ```powershell
 npm run build            # the script imports the compiled provider
-npm run sync-models      # fetch catalog, probe vision, merge into ~/.config/opencode/opencode.json
+npm run sync-models      # fetch catalog, pick models, probe vision, write ~/.config/opencode/opencode.json
 ```
 
+Keys: `↑/↓` move (one model per step), `space` toggles, `Ctrl-A` selects all, `Ctrl-U` clears,
+typing filters, `Esc` clears the filter (empty filter aborts), `enter` confirms, `Ctrl-C` aborts
+without writing. Each row shows the model id and name plus a second line with its context window
+(`ctx 1M (1000000)`); the catalog carries no pricing fields, so no cost is shown. Needs an
+interactive terminal.
+
+Before anything else the script requires a CommandCode API key (from
+`provider.commandcode.options.apiKey` or `COMMANDCODE_API_KEY`). If neither is present — including
+when `opencode.json` does not exist yet — it exits and points you at the CommandCode Console
+(<https://commandcode.ai/studio/provider>) to generate one. It never writes without a key.
+
 Then restart opencode and confirm:
 
 ```powershell
@@ -108,21 +121,11 @@ Mapping and behaviour:
   `qwen…vl`, …). Probing is best-effort: upstream is nondeterministic and a re-run may flip a
   borderline model.
 - Generated fields (`name`, `limit`, `reasoning`, `attachment`, `modalities`) overwrite existing
-  values. Other fields (`variants`, `options`, `cost`, …) and ids not in the catalog are preserved.
-
-Options:
-
-```powershell
-node scripts/sync-models.mjs --dry-run        # print the merged map, write nothing
-node scripts/sync-models.mjs --out models.json # write only the models fragment
-node scripts/sync-models.mjs --no-probe       # skip probing, use the heuristic (fast/offline)
-node scripts/sync-models.mjs --concurrency 8  # probe parallelism (default 6)
-node scripts/sync-models.mjs --no-preserve    # snapshot only: drop user ids/extra fields
-node scripts/sync-models.mjs --config <path>  # target a different config
-```
+  values. Other fields (`variants`, `options`, `cost`, …) on selected ids are preserved.
+  Unselected ids are removed from the map.
 
 The API key is read from `COMMANDCODE_API_KEY`, then `provider.commandcode.options.apiKey` in the
-target config.
+global config (`baseURL`/`ccVersion` are read from there too).
 
 
 ## Configuration options
@@ -143,26 +146,131 @@ target config.
 `options.headers` (with a leading `Bearer ` stripped). If neither is present, requests are sent
 unauthenticated and CommandCode will reject them.
 
+## Environment variables
+
+All environment variables read by the provider, its scripts, and its TUI plugin. String values in
+`options` also accept `{env:NAME}`, which opencode/the TUI expand generically (not a fixed set).
+
+CommandCode-specific:
+
+| Name | Purpose | Default |
+| --- | --- | --- |
+| `COMMANDCODE_API_KEY` | CommandCode API key for `smoke`/`quota`/`sync-models` and the TUI's key resolution; referenced from config as `{env:COMMANDCODE_API_KEY}`. | — |
+| `COMMANDCODE_TOGGLES_FILE` | Overrides the shared toggle-file path. | `$XDG_CONFIG_HOME/opencode/commandcode-toggles.json` |
+| `COMMANDCODE_QUOTA_INTERVAL_MS` | Sidebar fallback quota refresh interval in ms. | `300000` (floor `120000`) |
+
+Standard:
+
+| Name | Purpose | Default |
+| --- | --- | --- |
+| `XDG_CONFIG_HOME` | Base config dir for the toggle file. | `~/.config` |
+| `NO_COLOR` | Disables ANSI colors in the `sync-models` picker (any value). | unset |
+| `TERM` | `dumb` disables ANSI colors in the `sync-models` picker. | — |
+
 ## Debug tracing
 
-Set `COMMANDCODE_DEBUG=1` to write a trace of every request and stream event to a log file.
-Silent (and no file is created) unless enabled. Every line is passed through `redact()` so
+Flip the `debug` key in the shared toggle file (via `/cc-debug`) to write a trace of every
+request and stream event to a log file. Silent (and no file is created) unless enabled; the log
+path is `<os tempdir>/commandcode-debug.log`. Every line is passed through `redact()` so
 credentials never reach disk.
 
-```powershell
-$env:COMMANDCODE_DEBUG = "1"
-opencode run "Reply with exactly: pong" -m commandcode/deepseek/deepseek-v4.1-flash
-```
-
-| Env var | Default | Description |
-| --- | --- | --- |
-| `COMMANDCODE_DEBUG` | unset | `1`/`true`/`yes` enables tracing. |
-| `COMMANDCODE_DEBUG_FILE` | `<os tempdir>/commandcode-debug.log` | Where the trace is appended. |
-
 The log captures the request (model id, body byte length), each HTTP attempt (status, retry
 wait, elapsed ms), every stream event payload, and the terminal finish reason + usage. It is
 useful for diagnosing model selection, retry, tool-call, and finish-reason issues.
 
+## ZDR header toggle
+
+The `x-cmd-zdr: 1` request header is **omitted by default** because it is rejected by some
+models. It can be enabled with the `/cc-zdr` slash command or
+`providerOptions.commandcode.zdr`. The value is resolved per request, so toggling it
+does not require restarting opencode.
+
+Precedence, highest first:
+
+1. `providerOptions.commandcode.zdr` (`true`/`false`), e.g. a model variant or agent option.
+2. An explicit `x-cmd-zdr` header in `opencode.json` `options.headers` or the call's `headers`.
+3. The toggle file (`zdr` key — see [Runtime toggles](#runtime-toggles)).
+
+With tracing enabled, the trace logs a `zdr on|off` line per request so the toggle state is
+observable.
+
+## Runtime toggles
+
+`/cc-zdr` and `/cc-debug` flip the `zdr` and `debug` values in a small JSON file the provider
+reads on every request, so both settings change without restarting opencode. `/cc-status` shows
+the current state. The file is `~/.config/opencode/commandcode-toggles.json` (overridable with
+`COMMANDCODE_TOGGLES_FILE` — see [Environment variables](#environment-variables)).
+
+These commands come from a small TUI plugin shipped in this package (`src/tui.tsx`). It is loaded
+from **source** — opencode compiles the TSX and provides the Solid runtime itself — so there is no
+build step and no runtime dependency for it. Register it once by adding the plugin file's absolute
+`file://` URL to the `plugin` list in `tui.json` (global `~/.config/opencode/tui.json`, or
+project-local `.opencode/tui.json`):
+
+```jsonc
+// tui.json
+{
+  "plugin": ["file:///C:/dev/opencode-commandcode-provider/src/tui.tsx"]
+}
+```
+
+Use an absolute path with forward slashes (`file:///C:/...` on Windows, `file:///home/...` on
+Linux/macOS). Two rules for the entry:
+
+- It must point at **`src/tui.tsx` itself**. `opencode plugin <module>` cannot install this form:
+  the CLI requires a `package.json` manifest next to the target, so a bare `.tsx` must be wired in
+  by hand as above.
+- Never point `plugin` at the **repo root**. A directory spec makes opencode also treat the package
+  as a server plugin (via `package.json` `main` → `dist/index.js`), which is the AI SDK provider,
+  not a plugin — it loads as dead weight into every session.
+
+Restart opencode after registering the plugin; then type `/cc-` for autocomplete. The plugin also
+renders a **CommandCode panel in the session sidebar** (`zdr` / `debug`, plus live quota, above the
+built-in panels) that updates live as you toggle. Note that in opencode 1.x the slash menu and the
+`Ctrl+P` palette read the same command registry, so these entries appear in both. If you run opencode
+against a remote server, the TUI-side file is not visible to the provider on the server host — use
+the env vars or `providerOptions` there.
+
+## Quota tracking
+
+`/cc-usage` and the sidebar show CommandCode's **5-hour**, **weekly** and **monthly** usage for the
+account behind `provider.commandcode.options.apiKey`. The numbers come live from the same alpha
+billing endpoints the `cmd` CLI `/usage` command uses:
+
+| Endpoint | Used for |
+| --- | --- |
+| `GET /alpha/whoami` | account identity and org id |
+| `GET /alpha/billing/credits` | 5-hour/weekly windows (`used`, `cap`, `resetAt`) and credit balances |
+| `GET /alpha/billing/subscriptions` | plan id, status, billing period |
+| `GET /alpha/usage/summary` | requests/tokens/cost for the current billing period |
+
+The **5-hour** and **weekly** windows show `used / cap` and a live reset countdown. The **monthly**
+meter is derived: `usage/summary` cost vs. that cost plus all remaining credits, so it works on any
+plan (including org and pay-as-you-go) without a maintained price table. Free and pay-as-you-go
+accounts can return no `windowLimits`; those rows stay hidden and the missing section is reported
+instead of being shown as zero.
+
+The sidebar refreshes after each completed turn (`session.idle`, plus `session.status`,
+`session.updated`, and `message.updated` as fallbacks since `session.idle` is a server-plugin
+event that may never reach the TUI bus — all debounced) and when the active session changes,
+but never more than once every 2 minutes; every 5 minutes as a fallback
+(`COMMANDCODE_QUOTA_INTERVAL_MS` overrides the fallback interval — see
+[Environment variables](#environment-variables)) and the
+countdown ticks every 30 seconds. Key resolution and fetch failures are redacted and shown
+inline; the panel never stays on `loading…` and never blocks the provider. Extra trace lines
+(`tui-quota`) are appended to the debug log only when debug tracing is on.
+
+Headless check (no TUI), reading the key from `COMMANDCODE_API_KEY` or the opencode config:
+
+```powershell
+npm run build
+npm run quota            # formatted
+npm run quota -- --json  # raw quota object
+```
+
+The API key is never logged; quota errors pass through `redact()` like every other surface.
+
+
 ## Features
 
 | Capability | Status |
@@ -176,11 +284,15 @@ useful for diagnosing model selection, retry, tool-call, and finish-reason issue
 | Tool results | Yes — paired results are replayed; unpaired ids are dropped |
 | Multiple images (vision) | Yes — `data:` URIs, raw base64, `Uint8Array`, and remote URLs |
 | Token usage | Yes — input/output totals, cache read, reasoning tokens |
+| Provider-reported cost | Yes — per-request USD from `finish-step`/`provider-metadata` is forwarded as `providerMetadata.commandcode.cost` (plus `marketCost`); also stashed in `usage.raw`. The built-in sidebar still shows `$0.00` until opencode itself consumes this field (upstream `anomalyco/opencode#43818`); `/cc-usage` remains the accurate dollar source |
 | Finish reasons | Yes — unified (`stop`, `length`, `tool-calls`, `content-filter`, `error`, `other`) plus raw |
 | Sampling parameters | Yes — `temperature`, `topP`, `topK`, `stopSequences`, `seed`, presence/frequency penalties |
 | `reasoning_effort` | Yes — via `providerOptions.commandcode` |
-| Retry with backoff | Yes — 429/5xx and network errors, honouring `Retry-After` |
+| Retry with backoff | Yes — 429/5xx and network errors, honouring `Retry-After`; also retryable `error` events that arrive inside an HTTP 200 stream before any content (gateway 520s) |
 | Credential redaction | Yes — error bodies are scrubbed before surfacing |
+| Runtime toggles | Yes — `/cc-zdr` and `/cc-debug` flip the shared toggle file without a restart |
+| Sidebar status panel | Yes — a `sidebar_content` panel shows `zdr`/`debug` and live quota |
+| Quota tracking | Yes — 5-hour/weekly/monthly meters via `/cc-usage`, the sidebar, and `npm run quota` |
 
 ### `tool_choice` handling
 
@@ -213,10 +325,14 @@ src/transform.ts  LanguageModelV3CallOptions -> /alpha/generate envelope (JSON s
 src/events.ts     NDJSON/SSE line iterator over the upstream response body.
 src/usage.ts      finish event -> V3 usage; finish-reason unification.
 src/redact.ts     Credential scrubbing for error surfaces.
-src/log.ts        Opt-in tracing (COMMANDCODE_DEBUG) to a log file.
+src/log.ts        Opt-in tracing (toggle file) to a log file.
+src/toggles.ts    Shared toggle file read by the provider and written by the TUI plugin.
+src/quota.ts      Live 5-hour/weekly/monthly quota from the alpha billing endpoints.
+src/tui.tsx       opencode TUI plugin: /cc-zdr, /cc-debug, /cc-status, /cc-usage + sidebar panel.
 src/constants.ts  Defaults, header names, passthrough params, static config block.
 scripts/smoke.mjs Live end-to-end check against api.commandcode.ai.
-scripts/sync-models.mjs Generate provider.commandcode.models from the live catalog.
+scripts/quota.mjs Print the live quota headlessly.
+scripts/sync-models.mjs Pick models interactively and write provider.commandcode.models.
 ```
 
 ### Request lifecycle
@@ -241,9 +357,10 @@ scripts/sync-models.mjs Generate provider.commandcode.models from the live catal
 
 ```powershell
 npm install        # installs typescript, @types/node, @ai-sdk/provider
-npm run typecheck  # tsc --noEmit
-npm run build      # tsc -> dist/
+npm run typecheck  # tsc --noEmit (provider) + tsc -p tsconfig.tui.json (TUI plugin)
+npm run build      # tsc -> dist/ (provider only; the TUI plugin loads from source)
 npm run smoke      # live request against CommandCode
+npm run quota      # live 5-hour/weekly/monthly quota
 npm run sync-models # regenerate provider.commandcode.models from the catalog
 ```
 
@@ -274,9 +391,10 @@ opencode run "Use the glob tool to list *.mjs and report the filenames." -m comm
 | `stream ended ... no finish` / truncated output | Upstream closed the connection early. The provider emits a synthetic `finish`, but the response is incomplete; retry the turn. |
 | Images are ignored by the model | The selected model is not vision-capable. Mark it with `"attachment": true` and `modalities.input: ["text","image"]` in `models`, and pick a vision model id. |
 | Upstream 400 about tool calls | A `tool-call` or `tool-result` without a matching pair slipped through. Pairing is enforced in `src/transform.ts`; report a repro if it still occurs. |
-| Need to see what the provider sends/receives | Set `COMMANDCODE_DEBUG=1` and read the appended log file (see [Debug tracing](#debug-tracing)). |
+| Need to see what the provider sends/receives | Flip `/cc-debug` and read the appended log file (see [Debug tracing](#debug-tracing)). |
 | `tool_choice` seemingly ignored | Expected for `required`; upstream cannot force a call. `none` and named-tool are emulated via the tool list. |
 | Config change had no effect | opencode reads config once at startup. Restart it. |
+| `/cc-*` commands and the sidebar panel are missing | TUI plugin not registered: add a direct `file://` URL to `src/tui.tsx` in `tui.json` (see [Runtime toggles](#runtime-toggles)). A repo-root or `src/tui.tsx`-less entry does not load. |
 
 ## Security
 

package.json

@@ -12,12 +12,19 @@
     }
   },
   "files": [
-    "dist"
+    "dist",
+    "src/tui.tsx",
+    "src/tui-shims.d.ts",
+    "src/toggles.ts",
+    "src/quota.ts",
+    "src/redact.ts",
+    "src/constants.ts"
   ],
   "scripts": {
     "build": "tsc",
-    "typecheck": "tsc --noEmit",
+    "typecheck": "tsc --noEmit && tsc -p tsconfig.tui.json",
     "smoke": "node scripts/smoke.mjs",
+    "quota": "node scripts/quota.mjs",
     "sync-models": "node scripts/sync-models.mjs"
   },
   "devDependencies": {

scripts/quota.mjs

@@ -0,0 +1,91 @@
+// Print the live CommandCode quota (5-hour, weekly, monthly) from the alpha billing
+// endpoints. The same data drives the TUI sidebar and /cc-usage command; this script is
+// the headless check and does not need opencode running.
+//
+// Run: node scripts/quota.mjs [--config <path>] [--base-url <url>] [--json]   (build first: npm run build)
+import { readFileSync } from "node:fs";
+import { homedir } from "node:os";
+import { join } from "node:path";
+import { fetchQuota, formatQuota } from "../dist/quota.js";
+
+function parseArgs(argv) {
+  const opts = {
+    config: join(homedir(), ".config", "opencode", "opencode.json"),
+    baseURL: undefined,
+    json: false,
+  };
+  for (let i = 0; i < argv.length; i++) {
+    const arg = argv[i];
+    const value = () => {
+      const next = argv[++i];
+      if (next === undefined) throw new Error(`Missing value for ${arg}`);
+      return next;
+    };
+    switch (arg) {
+      case "--config":
+        opts.config = value();
+        break;
+      case "--base-url":
+        opts.baseURL = value();
+        break;
+      case "--json":
+        opts.json = true;
+        break;
+      case "-h":
+      case "--help":
+        console.log(
+          [
+            "Print CommandCode quota from the alpha billing endpoints.",
+            "",
+            "Usage: node scripts/quota.mjs [options]",
+            "  --config <path>   opencode config to read (default ~/.config/opencode/opencode.json)",
+            "  --base-url <url>  upstream origin (default from config or api.commandcode.ai)",
+            "  --json            print the raw quota object instead of formatted text",
+          ].join("\n"),
+        );
+        process.exit(0);
+        break;
+      default:
+        throw new Error(`Unknown argument: ${arg}`);
+    }
+  }
+  return opts;
+}
+
+function readConfig(path) {
+  try {
+    return JSON.parse(readFileSync(path, "utf8"));
+  } catch {
+    return {};
+  }
+}
+
+function resolveKey(config, env) {
+  const fromEnv = env.COMMANDCODE_API_KEY;
+  if (fromEnv && fromEnv.trim()) return fromEnv.trim();
+  const options = config?.provider?.commandcode?.options ?? {};
+  const fromConfig = options.apiKey;
+  if (typeof fromConfig === "string" && fromConfig.trim()) return fromConfig.trim();
+  const auth = options.headers?.Authorization ?? options.headers?.authorization;
+  if (typeof auth === "string" && auth.trim()) return auth.replace(/^Bearer\s+/i, "").trim();
+  return undefined;
+}
+
+const opts = parseArgs(process.argv.slice(2));
+const config = readConfig(opts.config);
+const apiKey = resolveKey(config, process.env);
+if (!apiKey) {
+  console.error(
+    `No CommandCode API key found. Set COMMANDCODE_API_KEY or provider.commandcode.options.apiKey in ${opts.config}.`,
+  );
+  process.exit(1);
+}
+
+const baseURL = opts.baseURL ?? config?.provider?.commandcode?.options?.baseURL;
+const result = await fetchQuota({ apiKey, baseURL: typeof baseURL === "string" ? baseURL : undefined });
+if (!result.ok) {
+  console.error(result.error.message);
+  process.exit(1);
+}
+
+console.log(opts.json ? JSON.stringify(result.quota, null, 2) : formatQuota(result.quota));

scripts/sync-models.mjs

@@ -1,23 +1,29 @@
-// Generate the opencode `provider.commandcode.models` map from CommandCode's live catalog.
+// Interactive model picker for opencode `provider.commandcode.models`.
 //
 // opencode builds `/models` from the configured model map; a custom `file://` provider is only
 // ever asked for `languageModel(id)`, never queried for discovery. So the catalog has to be baked
-// into opencode.json. This script fetches it, probes each model for vision support, and merges the
-// result into the target config.
+// into opencode.json. This script fetches the live catalog, lets the user pick which models to
+// enable with a builtins-only checkbox UI, probes the selection for vision support, and writes
+// the result into the global opencode config. Deselecting a model removes it from the config.
 //
-// Run: node scripts/sync-models.mjs [options]   (build first: npm run build)
-import { readFileSync, writeFileSync, renameSync } from "node:fs";
+// Run: node scripts/sync-models.mjs   (build first: npm run build)
+import { mkdirSync, readFileSync, writeFileSync, renameSync } from "node:fs";
 import { homedir } from "node:os";
-import { join } from "node:path";
+import { dirname, join } from "node:path";
+import readline from "node:readline";
 import { createCommandCode } from "../dist/index.js";
 import { redact } from "../dist/redact.js";
 import { DEFAULT_BASE_URL, DEFAULT_CC_VERSION, DEFAULT_MAX_TOKENS, MODELS_PATH } from "../dist/constants.js";
 
+const CONFIG_PATH = join(homedir(), ".config", "opencode", "opencode.json");
+const CONSOLE_URL = "https://commandcode.ai/studio/provider";
+const CONCURRENCY = 6;
+
 // 64x64 solid red PNG, used only to provoke a vision-accepting vs vision-rejecting response.
 const PROBE_IMAGE = {
   type: "file",
   mediaType: "image/png",
-  data: "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAEAAAABACAIAAAAlC+aJAAAAeUlEQVR4nO3PQQkAMAzAwCqpf1ETMxF7HINABFzm7H7dcEEDWtCAFjSgBQ1oQQNa0IAWNKAFDWhBA1rQgBY0oAUNaEEDWtCAFjSgBQ1oQQNa0IAWNKAFDWhBA1rQgBY0oAUNaEEDWtCAFjSgBQ1oQQNa0IAWNKAFj13PLIEAOXyUUwAAAABJRU5ErkJggg==",
+  data: "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAEAAAABACAIAAAAlC+aJAAAAeUlEQVR4nO3PQQkAMAzAwCqpf1ETMxF7HINABFzm7H7dcEEDWtCAFjSgBQ1oQQNa0IAWNKAFDWhBA1rQgBY0oAUNaEEDWtCAFjSgBQ1oQQNa0IAWNKAFj13PLIEAOXyUUwAAAABJRU5ErkJggg==",
 };
 
 const VISION_ERROR = /image|vision|multimodal|modality|unsupported|not support|invalid.*content|does not support/i;
@@ -26,63 +32,20 @@ const VISION_ERROR = /image|vision|multimodal|modality|unsupported|not support|i
 const NO_IMAGE =
   /(?:don'?t|do not|can'?t|cannot|unable to)\s+(?:see|view|detect|find|access|receive|open|process)|no\s+(?:image|picture|attachment|photo)|not\s+(?:see|receive|detect|attach)|image\s+(?:was\s+)?not\s+(?:attach|provid|receiv|includ)/i;
 
-// Fallback when a probe is inconclusive or disabled. Ground truth is unavailable from the catalog,
+// Fallback when a probe is inconclusive. Ground truth is unavailable from the catalog,
 // so this is only a best guess for ids that look like known vision families.
 const VISION_HEURISTIC = /claude|gpt-5|gemini|grok|qwen.*vl|vision|(?:^|[^a-z])vl(?:[^a-z]|$)|omni|multimodal/i;
 
-function parseArgs(argv) {
-  const opts = {
-    config: join(homedir(), ".config", "opencode", "opencode.json"),
-    baseURL: undefined,
-    out: undefined,
-    dryRun: false,
-    probe: true,
-    concurrency: 6,
-    preserve: true,
-  };
-  for (let i = 0; i < argv.length; i++) {
-    const arg = argv[i];
-    const value = () => {
-      const next = argv[++i];
-      if (next === undefined) throw new Error(`Missing value for ${arg}`);
-      return next;
-    };
-    switch (arg) {
-      case "--config": opts.config = value(); break;
-      case "--base-url": opts.baseURL = value(); break;
-      case "--out": opts.out = value(); break;
-      case "--concurrency": opts.concurrency = Math.max(1, Number(value()) || 1); break;
-      case "--dry-run": opts.dryRun = true; break;
-      case "--no-probe": opts.probe = false; break;
-      case "--no-preserve": opts.preserve = false; break;
-      case "-h":
-      case "--help":
-        console.log(
-          [
-            "Generate provider.commandcode.models from the CommandCode catalog.",
-            "",
-            "Usage: node scripts/sync-models.mjs [options]",
-            "  --config <path>     opencode config to update (default ~/.config/opencode/opencode.json)",
-            "  --base-url <url>    upstream origin (default from config or api.commandcode.ai)",
-            "  --out <file>        write only the models fragment; do not touch the config",
-            "  --dry-run           print the merged map; write nothing",
-            "  --no-probe          skip vision probing; use the family heuristic",
-            "  --concurrency <n>   vision probe parallelism (default 6)",
-            "  --no-preserve       snapshot only; drop user-added ids and extra fields",
-          ].join("\n"),
-        );
-        process.exit(0);
-        break;
-      default:
-        throw new Error(`Unknown argument: ${arg}`);
-    }
-  }
-  return opts;
-}
-
 function readConfig(path) {
+  try {
     const text = readFileSync(path, "utf8");
-  return { text, data: JSON.parse(text) };
+    return { text, data: JSON.parse(text), exists: true };
+  } catch (error) {
+    // A fresh install has no opencode.json yet; treat it as empty so the console
+    // guidance / env-key path can still run. Bad JSON or unreadable files still throw.
+    if (error.code === "ENOENT") return { text: "", data: {}, exists: false };
+    throw error;
+  }
 }
 
 function detectIndent(text) {
@@ -197,34 +160,248 @@ function modelEntry(catalogModel, vision) {
   return entry;
 }
 
-function mergeModels(existing, generated, preserve) {
-  const out = {};
-  if (preserve) {
-    for (const [id, model] of Object.entries(existing ?? {})) out[id] = model;
-  }
-  for (const [id, entry] of Object.entries(generated)) {
-    out[id] = preserve ? { ...(existing?.[id] ?? {}), ...entry } : entry;
-  }
-  return out;
-}
-
 function writeAtomic(path, contents) {
   const tmp = `${path}.${process.pid}.tmp`;
   writeFileSync(tmp, contents);
   renameSync(tmp, path);
 }
 
-async function main() {
-  const opts = parseArgs(process.argv.slice(2));
-  const { text, data: config } = readConfig(opts.config);
+// The catalog carries no pricing fields (id/name/context_length/supported_endpoints only),
+// so the picker shows context with any opportunistically present pricing keys omitted.
+function formatContext(value) {
+  const n = Number(value);
+  if (!Number.isFinite(n) || n <= 0) return "unknown";
+  if (n >= 1_000_000 && n % 1_000_000 === 0) return `${n / 1_000_000}M (${n})`;
+  if (n >= 1_000_000) return `${(n / 1_000_000).toFixed(1)}M (${n})`;
+  if (n >= 1000 && n % 1000 === 0) return `${n / 1000}k (${n})`;
+  if (n >= 1000) return `${(n / 1000).toFixed(0)}k (${n})`;
+  return `${n}`;
+}
 
-  const key = resolveKey(config, process.env);
-  if (!key) {
-    console.error("No API key: set COMMANDCODE_API_KEY or provider.commandcode.options.apiKey");
+// Builtins-only checkbox picker (no dependencies, so the package stays runtime-dependency-free).
+// Nothing is pre-checked; deselected ids are dropped by the caller. Resolves to the selected
+// catalog indices, or null when the user aborts (Esc on empty filter / Ctrl-C).
+function pickModels(catalog, existingIds) {
+  return new Promise((resolve) => {
+    const checked = new Set();
+    let filter = "";
+    let cursor = 0;
+    let scroll = 0;
+    let done = false;
+
+    const columns = () => process.stdout.columns ?? 80;
+    // Two terminal lines per model (header + context detail); keep at least one model visible.
+    const pageSize = () => Math.max(1, Math.floor(Math.max(5, (process.stdout.rows ?? 24) - 7) / 2));
+
+    const filtered = () => {
+      if (!filter) return catalog.map((_, index) => index);
+      const needle = filter.toLowerCase();
+      const out = [];
+      catalog.forEach((model, index) => {
+        if (`${model.id} ${model.name ?? ""}`.toLowerCase().includes(needle)) out.push(index);
+      });
+      return out;
+    };
+
+    const clamp = (list) => {
+      if (list.length === 0) {
+        cursor = 0;
+        scroll = 0;
+        return;
+      }
+      cursor = Math.min(Math.max(0, cursor), list.length - 1);
+      const size = pageSize();
+      if (cursor < scroll) scroll = cursor;
+      if (cursor >= scroll + size) scroll = cursor - size + 1;
+    };
+
+    const line = (text, code) => {
+      // Slice visible text first so ANSI codes never count toward the terminal
+      // width or get cut mid-sequence. NO_COLOR and dumb terminals get plain text.
+      const visible = String(text).slice(0, columns());
+      if (code === undefined || process.env.NO_COLOR || process.env.TERM === "dumb") {
+        process.stdout.write(`${visible}\n`);
+        return;
+      }
+      process.stdout.write(`\x1b[${code}m${visible}\x1b[0m\n`);
+    };
+
+    const render = () => {
+      const list = filtered();
+      clamp(list);
+      const size = pageSize();
+      const view = list.slice(scroll, scroll + size);
+      process.stdout.write("\x1b[?25l\x1b[2J\x1b[H");
+      line("Select CommandCode models  (type to filter · space toggle · enter confirm)");
+      line(`Catalog: ${catalog.length} models · Selected: ${checked.size} · Filter: ${filter}\u2588`);
+      line("");
+      if (list.length === 0) {
+        line("  (no matches)");
+      }
+      view.forEach((original, i) => {
+        const model = catalog[original];
+        const active = scroll + i === cursor;
+        const selected = checked.has(original);
+        const mark = selected ? "x" : " ";
+        const current = existingIds.has(model.id) ? " · current" : "";
+        // Selected rows render green (both lines); the cursor row renders bold on top.
+        const code = selected ? (active ? "1;32" : "32") : active ? "1" : undefined;
+        line(`${active ? "\u203a" : " "} [${mark}] ${model.id} \u2014 ${model.name ?? model.id}${current}`, code);
+        line(`    ctx ${formatContext(model.context_length)}`, code);
+      });
+      line("");
+      line("\u2191/\u2193 navigate · space toggle · ctrl-a all · ctrl-u none · esc back/abort · enter confirm");
+    };
+
+    const cleanup = () => {
+      if (done) return;
+      done = true;
+      process.stdout.removeListener("resize", render);
+      process.stdin.removeListener("keypress", onKey);
+      try {
+        process.stdin.setRawMode(false);
+      } catch {
+        // Non-TTY teardown; terminal state is already sane.
+      }
+      process.stdin.pause();
+      process.stdout.write("\x1b[?25h");
+    };
+
+    const finish = (value) => {
+      cleanup();
+      resolve(value);
+    };
+
+    const toggleAll = (list, value) => {
+      for (const index of list) {
+        if (value) checked.add(index);
+        else checked.delete(index);
+      }
+    };
+
+    function onKey(str, key) {
+      if (done) return;
+      const list = filtered();
+      if (key.ctrl && (key.name === "c" || key.name === "d")) {
+        finish(null);
+        return;
+      }
+      if (key.ctrl && key.name === "a") {
+        toggleAll(list, true);
+        render();
+        return;
+      }
+      if (key.ctrl && key.name === "u") {
+        checked.clear();
+        render();
+        return;
+      }
+      switch (key.name) {
+        case "return":
+          finish([...checked].sort((a, b) => a - b));
+          return;
+        case "escape":
+          // First Esc clears the filter; a second Esc on an empty filter aborts.
+          if (filter) {
+            filter = "";
+            cursor = 0;
+            scroll = 0;
+            render();
+            return;
+          }
+          finish(null);
+          return;
+        case "backspace":
+          if (filter) {
+            filter = filter.slice(0, -1);
+            cursor = 0;
+            scroll = 0;
+            render();
+          }
+          return;
+        case "up":
+          cursor = list.length === 0 ? 0 : (cursor - 1 + list.length) % list.length;
+          render();
+          return;
+        case "down":
+          cursor = list.length === 0 ? 0 : (cursor + 1) % list.length;
+          render();
+          return;
+        case "space": {
+          const current = list[cursor];
+          if (current !== undefined) {
+            if (checked.has(current)) checked.delete(current);
+            else checked.add(current);
+          }
+          render();
+          return;
+        }
+        default:
+          break;
+      }
+      // All printable characters feed the filter — letter shortcuts would swallow
+      // filter input (e.g. typing "qwen" would abort on the "q"). Bulk actions live
+      // on Ctrl combos above; abort is Esc (empty filter) or Ctrl-C.
+      if (!key.ctrl && !key.meta && typeof str === "string" && str.length === 1) {
+        const printable = str >= " " && str <= "~";
+        if (printable && str !== " ") {
+          filter += str;
+          cursor = 0;
+          scroll = 0;
+          render();
+        }
+      }
+    }
+
+    readline.emitKeypressEvents(process.stdin);
+    try {
+      process.stdin.setRawMode(true);
+    } catch {
+      resolve(null);
+      return;
+    }
+    process.stdin.resume();
+    process.stdout.on("resize", render);
+    process.stdin.on("keypress", onKey);
+    render();
+  });
+}
+
+async function main() {
+  let config;
+  let text;
+  let exists;
+  try {
+    ({ text, data: config, exists } = readConfig(CONFIG_PATH));
+  } catch (error) {
+    console.error(`Cannot read ${CONFIG_PATH}: ${error instanceof Error ? error.message : String(error)}`);
     process.exit(1);
   }
 
-  const baseURL = (opts.baseURL ?? config?.provider?.commandcode?.options?.baseURL ?? DEFAULT_BASE_URL).replace(/\/+$/, "");
+  // Mandatory first step: without a CommandCode key there is nothing to fetch or
+  // probe, so point a fresh user at the Console before doing anything else.
+  const key = resolveKey(config, process.env);
+  if (!key) {
+    const where = exists
+      ? `No CommandCode API key in ${CONFIG_PATH}.`
+      : `No opencode.json found at ${CONFIG_PATH}.`;
+    console.error(
+      `${where}\n` +
+        `Generate an API key in the CommandCode Console: ${CONSOLE_URL}\n` +
+        (exists
+          ? `Add it as provider.commandcode.options.apiKey, or export COMMANDCODE_API_KEY.`
+          : `Create that file with provider.commandcode.options.apiKey, or export COMMANDCODE_API_KEY.`) +
+        `\nNothing was written.`,
+    );
+    process.exit(1);
+  }
+
+  if (!process.stdin.isTTY || !process.stdout.isTTY) {
+    console.error("Interactive picker needs a TTY; nothing was written.");
+    process.exit(1);
+  }
+
+  const baseURL = (config?.provider?.commandcode?.options?.baseURL ?? DEFAULT_BASE_URL).replace(/\/+$/, "");
   const ccVersion = config?.provider?.commandcode?.options?.ccVersion ?? DEFAULT_CC_VERSION;
 
   console.log(`Fetching catalog: ${baseURL}${MODELS_PATH}`);
@@ -235,24 +412,40 @@ async function main() {
     console.error(`Failed to fetch catalog: ${error instanceof Error ? error.message : String(error)}`);
     process.exit(1);
   }
+  if (catalog.length === 0) {
+    console.error("Catalog is empty; config unchanged.");
+    process.exit(1);
+  }
   console.log(`Catalog: ${catalog.length} models`);
 
-  const provider = createCommandCode({ name: "commandcode", apiKey: key, baseURL, ccVersion, maxRetries: 1 });
+  const existing = config?.provider?.commandcode?.models ?? {};
+  const existingIds = new Set(Object.keys(existing));
 
-  let probeResults = [];
-  if (opts.probe) {
-    console.log(`Probing vision (${catalog.length} models, concurrency ${opts.concurrency})...`);
-    probeResults = await mapPool(catalog, opts.concurrency, async (model, index) => {
-      const outcome = await probeVisionConfirmed(provider, model.id);
-      const tag = outcome.vision === true ? "vision" : outcome.vision === false ? "no-vision" : "unknown";
-      console.log(`  [${index + 1}/${catalog.length}] ${model.id}: ${tag} (${outcome.reason})`);
-      return outcome;
-    });
+  const picked = await pickModels(catalog, existingIds);
+  if (picked === null) {
+    console.log("Aborted \u2014 config unchanged.");
+    return;
+  }
+  if (picked.length === 0) {
+    console.log("No models selected \u2014 config unchanged.");
+    return;
   }
 
-  const generated = {};
+  const selected = picked.map((index) => catalog[index]).filter(Boolean);
+  console.log(`Probing vision (${selected.length} models, concurrency ${CONCURRENCY})...`);
+  const provider = createCommandCode({ name: "commandcode", apiKey: key, baseURL, ccVersion, maxRetries: 1 });
+  const probeResults = await mapPool(selected, CONCURRENCY, async (model, index) => {
+    const outcome = await probeVisionConfirmed(provider, model.id);
+    const tag = outcome.vision === true ? "vision" : outcome.vision === false ? "no-vision" : "unknown";
+    console.log(`  [${index + 1}/${selected.length}] ${model.id}: ${tag} (${outcome.reason})`);
+    return outcome;
+  });
+
+  // Deselect = remove: the final map holds exactly the picked ids. Extra user fields on
+  // picked ids are preserved; unpicked ids are dropped.
+  const merged = {};
   const stats = { probed: 0, vision: 0, noVision: 0, unknown: 0, heuristic: 0 };
-  catalog.forEach((model, index) => {
+  selected.forEach((model, index) => {
     const outcome = probeResults[index];
     let vision;
     if (outcome && outcome.vision !== undefined) {
@@ -265,41 +458,24 @@ async function main() {
       stats.unknown++;
       if (vision) stats.heuristic++;
     }
-    generated[model.id] = modelEntry(model, vision);
+    merged[model.id] = { ...(existing[model.id] ?? {}), ...modelEntry(model, vision) };
   });
 
-  const existing = config?.provider?.commandcode?.models ?? {};
-  const merged = mergeModels(existing, generated, opts.preserve);
-  const added = Object.keys(generated).filter((id) => !(id in existing)).length;
-  const preservedIds = Object.keys(existing).filter((id) => !(id in generated)).length;
-
-  if (opts.out) {
-    writeAtomic(opts.out, `${JSON.stringify(generated, null, 2)}\n`);
-    console.log(`Wrote ${Object.keys(generated).length} models to ${opts.out}`);
-    return;
-  }
-
-  if (opts.dryRun) {
-    console.log(JSON.stringify(merged, null, 2));
-    console.log(
-      `\n[dry-run] generated=${Object.keys(generated).length} added=${added} ` +
-        `preserved-ids=${preservedIds} probed=${stats.probed} vision=${stats.vision} ` +
-        `no-vision=${stats.noVision} unknown=${stats.unknown} heuristic-vision=${stats.heuristic}`,
-    );
-    return;
-  }
+  const removed = [...existingIds].filter((id) => !(id in merged)).length;
 
   config.provider ??= {};
   config.provider.commandcode ??= {};
   config.provider.commandcode.models = merged;
 
   const indent = detectIndent(text);
-  const trailing = text.endsWith("\n") ? "\n" : "";
-  writeAtomic(opts.config, `${JSON.stringify(config, null, indent)}${trailing}`);
+  const trailing = text === "" || text.endsWith("\n") ? "\n" : "";
+  // The config directory may not exist yet on a fresh install; create it before the atomic write.
+  mkdirSync(dirname(CONFIG_PATH), { recursive: true });
+  writeAtomic(CONFIG_PATH, `${JSON.stringify(config, null, indent)}${trailing}`);
 
   console.log(
-    `Wrote ${Object.keys(merged).length} models to ${opts.config}\n` +
-      `  generated=${Object.keys(generated).length} added=${added} preserved-ids=${preservedIds}\n` +
+    `Wrote ${Object.keys(merged).length} models to ${CONFIG_PATH}\n` +
+      `  selected=${selected.length} removed=${removed}\n` +
       `  probed=${stats.probed} vision=${stats.vision} no-vision=${stats.noVision} ` +
       `unknown=${stats.unknown} heuristic-vision=${stats.heuristic}\n` +
       `Restart opencode, then run: opencode models commandcode`,

src/events.ts

@@ -1,7 +1,10 @@
 /** CommandCode streams NDJSON; tolerate SSE-style `data:` prefixes and `[DONE]` sentinels. */
-import { debug } from "./log.js";
+import { debugWhen, isDebugEnabled } from "./log.js";
 
-export async function* iterateEvents(body: ReadableStream<Uint8Array>): AsyncGenerator<Record<string, any>> {
+export async function* iterateEvents(
+  body: ReadableStream<Uint8Array>,
+  dbg: boolean = isDebugEnabled(),
+): AsyncGenerator<Record<string, any>> {
   const reader = body.getReader();
   const decoder = new TextDecoder();
   let buffer = "";
@@ -16,7 +19,7 @@ export async function* iterateEvents(body: ReadableStream<Uint8Array>): AsyncGen
       while ((newline = buffer.indexOf("\n")) >= 0) {
         const raw = buffer.slice(0, newline);
         buffer = buffer.slice(newline + 1);
-        const event = parseLine(raw);
+        const event = parseLine(raw, dbg);
         if (event) yield event;
       }
     }
@@ -24,7 +27,7 @@ export async function* iterateEvents(body: ReadableStream<Uint8Array>): AsyncGen
     // flush any trailing line without a newline
     buffer += decoder.decode();
     if (buffer.trim()) {
-      const event = parseLine(buffer);
+      const event = parseLine(buffer, dbg);
       if (event) yield event;
     }
   } finally {
@@ -32,7 +35,7 @@ export async function* iterateEvents(body: ReadableStream<Uint8Array>): AsyncGen
   }
 }
 
-function parseLine(raw: string): Record<string, any> | null {
+function parseLine(raw: string, dbg: boolean): Record<string, any> | null {
   let line = raw.trim();
   if (!line || line.startsWith(":") || line.startsWith("event:")) return null;
   if (line.startsWith("data:")) line = line.slice(5).trim();
@@ -41,7 +44,7 @@ function parseLine(raw: string): Record<string, any> | null {
     const parsed = JSON.parse(line);
     return parsed && typeof parsed === "object" ? parsed : null;
   } catch {
-    debug("events", "parse-skip", `line=${line.slice(0, 512)}`);
+    debugWhen(dbg, "events", "parse-skip", `line=${line.slice(0, 512)}`);
     return null;
   }
 }

src/log.ts

@@ -1,13 +1,14 @@
-// Opt-in tracing to a log file. Gated on COMMANDCODE_DEBUG=1; silent (and no file
-// created) by default. Uses only Node built-ins so the built output stays dependency-free.
+// Opt-in tracing to a log file. Gated on the shared toggle file (debug key);
+// silent (and no file created) by default. Uses only Node built-ins so the built output
+// stays dependency-free.
 
 import { appendFileSync } from "node:fs";
 import { tmpdir } from "node:os";
 import { join } from "node:path";
 import { redact } from "./redact.js";
+import { debugEnabled } from "./toggles.js";
 
-const ENABLED = /^(1|true|yes)$/i.test(process.env["COMMANDCODE_DEBUG"] ?? "");
-const FILE = process.env["COMMANDCODE_DEBUG_FILE"] ?? join(tmpdir(), "commandcode-debug.log");
+const FILE = join(tmpdir(), "commandcode-debug.log");
 
 function serialize(value: unknown): string {
   if (typeof value === "string") return value;
@@ -18,9 +19,7 @@ function serialize(value: unknown): string {
   }
 }
 
-/** Write a trace line (appended) if tracing is enabled. Values are redacted before write. */
-export function debug(scope: string, ...args: unknown[]): void {
-  if (!ENABLED) return;
+function emit(scope: string, args: unknown[]): void {
   const parts = args.map(serialize);
   const line = `[commandcode] ${new Date().toISOString()} [${scope}] ${parts.join(" ")}`;
   try {
@@ -34,3 +33,20 @@ export function debug(scope: string, ...args: unknown[]): void {
     }
   }
 }
+
+/** Effective default tracing state (toggle file). Re-read on every call. */
+export function isDebugEnabled(): boolean {
+  return debugEnabled();
+}
+
+/** Write a trace line (appended) if tracing is enabled. Values are redacted before write. */
+export function debug(scope: string, ...args: unknown[]): void {
+  if (!debugEnabled()) return;
+  emit(scope, args);
+}
+
+/** Trace with a per-request override, so providerOptions/history state is not global. */
+export function debugWhen(enabled: boolean, scope: string, ...args: unknown[]): void {
+  if (!enabled) return;
+  emit(scope, args);
+}

src/model.ts

@@ -7,6 +7,7 @@ import type {
   LanguageModelV3StreamPart,
   LanguageModelV3StreamResult,
   LanguageModelV3Usage,
+  SharedV3ProviderMetadata,
   SharedV3Warning,
 } from "@ai-sdk/provider";
 import {
@@ -18,10 +19,38 @@ import {
   GENERATE_PATH,
 } from "./constants.js";
 import { iterateEvents } from "./events.js";
-import { debug } from "./log.js";
+import { debugWhen, isDebugEnabled } from "./log.js";
 import { redact } from "./redact.js";
+import { zdrEnabled } from "./toggles.js";
 import { transform } from "./transform.js";
-import { finishReasonFrom, usageFromFinish, type FinishEvent } from "./usage.js";
+import { finishReasonFrom, usageFromFinish, costFromFinishStep, costFromProviderMetadata, type FinishEvent, type FinishStepEvent, type ProviderMetadataEvent } from "./usage.js";
+
+// x-cmd-zdr breaks some models; off by default (server.py never sends it).
+// Precedence: providerOptions.commandcode.zdr > x-cmd-zdr header > toggle file.
+// "undefined" means the header is omitted entirely.
+function providerScoped(options: LanguageModelV3CallOptions): Record<string, unknown> | undefined {
+  return options.providerOptions?.["commandcode"] as Record<string, unknown> | undefined;
+}
+
+function truthyHeader(value: string): boolean {
+  return !/^(0|false|no|off)$/i.test(value);
+}
+
+function resolveZdr(
+  options: LanguageModelV3CallOptions,
+  providerHeaders: Record<string, string>,
+): string | undefined {
+  const scoped = providerScoped(options)?.["zdr"];
+  if (typeof scoped === "boolean") return scoped ? "1" : "0";
+  const header = headerValue({ ...providerHeaders, ...(options.headers ?? {}) }, "x-cmd-zdr");
+  if (header !== undefined) return truthyHeader(header) ? "1" : "0";
+  return zdrEnabled() ? "1" : undefined;
+}
+
+function resolveDebug(options: LanguageModelV3CallOptions): boolean {
+  const scoped = providerScoped(options)?.["debug"];
+  return typeof scoped === "boolean" ? scoped : isDebugEnabled();
+}
 
 export type CommandCodeOptions = {
   name?: string;
@@ -43,7 +72,7 @@ type ResolvedOptions = {
   retryMaxDelay: number;
 };
 
-function headerValue(headers: Record<string, string>, key: string): string | undefined {
+function headerValue(headers: Record<string, string | undefined>, key: string): string | undefined {
   const wanted = key.toLowerCase();
   for (const [k, v] of Object.entries(headers)) if (k.toLowerCase() === wanted) return v;
   return undefined;
@@ -80,6 +109,50 @@ function isRetryable(status: number): boolean {
   return status === 429 || (status >= 500 && status < 600);
 }
 
+// Upstream can answer HTTP 200 and then fail inside the stream (SSE `error` event).
+// CommandCode's gateway marks these with `statusCode`/`isRetryable`; retry before content.
+type StreamErrorInfo = { message: string; statusCode?: number; isRetryable?: boolean };
+
+function streamErrorInfo(evt: Record<string, any>): StreamErrorInfo | null {
+  if (evt?.["type"] !== "error") return null;
+  const err = evt["error"];
+  if (err === null || typeof err !== "object") return { message: String(err) };
+  const record = err as Record<string, unknown>;
+  const message = typeof record["message"] === "string" ? record["message"] : JSON.stringify(err);
+  return {
+    message,
+    ...(typeof record["statusCode"] === "number" ? { statusCode: record["statusCode"] } : {}),
+    ...(typeof record["isRetryable"] === "boolean" ? { isRetryable: record["isRetryable"] } : {}),
+  };
+}
+
+function isRetryableStreamError(info: StreamErrorInfo): boolean {
+  return info.isRetryable === true || (info.statusCode !== undefined && isRetryable(info.statusCode));
+}
+
+// Only `start`/`start-step` are safe to buffer while looking for an immediate stream error;
+// once real output appears a retry would duplicate the response.
+const STREAM_PREAMBLE = new Set(["start", "start-step"]);
+const MAX_PEEK_EVENTS = 32;
+
+type OpenedStream = {
+  response: Response;
+  prefix: Record<string, any>[];
+  events: AsyncGenerator<Record<string, any>>;
+};
+
+async function* emptyEvents(): AsyncGenerator<Record<string, any>> {
+  /* no-op */
+}
+
+async function* replayEvents(
+  prefix: Record<string, any>[],
+  rest: AsyncGenerator<Record<string, any>>,
+): AsyncGenerator<Record<string, any>> {
+  for (const evt of prefix) yield evt;
+  for await (const evt of rest) yield evt;
+}
+
 function retryDelay(attempt: number, retryAfter: string | null, maxDelay: number): number {
   if (retryAfter) {
     const seconds = Number(retryAfter);
@@ -131,7 +204,11 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
 
   supportedUrls: Record<string, RegExp[]> = {};
 
-  private requestHeaders(extra?: Record<string, string | undefined>): Record<string, string> {
+  private requestHeaders(
+    extra: Record<string, string | undefined> | undefined,
+    zdr: string | undefined,
+    dbg: boolean,
+  ): Record<string, string> {
     const headers: Record<string, string> = {
       "Content-Type": "application/json",
       "x-command-code-version": this.opts.ccVersion,
@@ -141,20 +218,26 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
       "x-co-flag": "false",
       ...this.opts.headers,
     };
+    if (zdr !== undefined) headers["x-cmd-zdr"] = zdr;
+    debugWhen(dbg, "fetch", "zdr", zdr === "1" ? "on" : "off");
     const auth = bearer(this.opts.apiKey);
     if (auth) headers["Authorization"] = auth;
     for (const [k, v] of Object.entries(extra ?? {})) if (v !== undefined) headers[k] = v;
     return headers;
   }
 
-  private async fetchWithRetry(body: string, options: LanguageModelV3CallOptions): Promise<Response> {
-    const headers = this.requestHeaders(options.headers);
+  private async fetchWithRetry(
+    body: string,
+    options: LanguageModelV3CallOptions,
+    dbg: boolean,
+  ): Promise<OpenedStream> {
+    const headers = this.requestHeaders(options.headers, resolveZdr(options, this.opts.headers), dbg);
     let lastError: unknown;
     const url = `${this.opts.baseURL}${GENERATE_PATH}`;
 
     for (let attempt = 0; attempt <= this.opts.maxRetries; attempt++) {
       if (options.abortSignal?.aborted) {
-        debug("fetch", "aborted", `attempt=${attempt}`);
+        debugWhen(dbg, "fetch", "aborted", `attempt=${attempt}`);
         throw options.abortSignal.reason ?? new Error("Aborted");
       }
       const started = Date.now();
@@ -169,7 +252,8 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
       } catch (error) {
         lastError = error;
         const wait = retryDelay(attempt, null, this.opts.retryMaxDelay);
-        debug(
+        debugWhen(
+          dbg,
           "fetch",
           "network-error",
           `url=${url}`,
@@ -188,7 +272,8 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
       if (isRetryable(status) && attempt < this.opts.maxRetries) {
         const wait = retryDelay(attempt, response.headers.get("retry-after"), this.opts.retryMaxDelay);
         if (wait >= 0) {
-          debug(
+          debugWhen(
+            dbg,
             "fetch",
             "retry",
             `url=${url}`,
@@ -206,14 +291,66 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
           continue;
         }
       }
-      debug("fetch", "response", `url=${url}`, `attempt=${attempt}`, `status=${status}`, `elapsedMs=${Date.now() - started}`);
-      return response;
+      if (!response.ok || !response.body) {
+        debugWhen(dbg, "fetch", "response", `url=${url}`, `attempt=${attempt}`, `status=${status}`, `elapsedMs=${Date.now() - started}`);
+        return { response, prefix: [], events: emptyEvents() };
+      }
+
+      // HTTP 200 can still hide an upstream failure. Buffer only the preamble; if the first
+      // real event is a retryable stream error, retry before anything reaches the client.
+      const events = iterateEvents(response.body, dbg);
+      const prefix: Record<string, any>[] = [];
+      let streamRetry: StreamErrorInfo | undefined;
+      while (prefix.length < MAX_PEEK_EVENTS) {
+        const next = await events.next();
+        if (next.done) break;
+        const evt = next.value;
+        prefix.push(evt);
+        const info = streamErrorInfo(evt);
+        if (info) {
+          if (isRetryableStreamError(info) && attempt < this.opts.maxRetries) streamRetry = info;
+          break;
+        }
+        if (!STREAM_PREAMBLE.has(evt["type"])) break;
+      }
+
+      if (streamRetry) {
+        const wait = retryDelay(attempt, null, this.opts.retryMaxDelay);
+        debugWhen(
+          dbg,
+          "fetch",
+          "stream-retry",
+          `url=${url}`,
+          `attempt=${attempt}`,
+          `status=${streamRetry.statusCode ?? "unknown"}`,
+          `message=${redact(streamRetry.message)}`,
+          `waitMs=${wait}`,
+        );
+        try {
+          await events.return(undefined as never);
+        } catch {
+          /* ignore */
+        }
+        try {
+          await response.body.cancel();
+        } catch {
+          /* ignore */
+        }
+        if (options.abortSignal?.aborted) {
+          throw options.abortSignal.reason ?? new Error("Aborted");
+        }
+        await sleep(wait);
+        continue;
+      }
+
+      debugWhen(dbg, "fetch", "response", `url=${url}`, `attempt=${attempt}`, `status=${status}`, `elapsedMs=${Date.now() - started}`);
+      return { response, prefix, events };
     }
 
     throw lastError instanceof Error ? lastError : new Error("Upstream unreachable");
   }
 
-  private async errorFrom(response: Response): Promise<Error> {
+  private async errorFrom(response: Response, dbg: boolean): Promise<Error> {
     let body = "";
     try {
       body = await response.text();
@@ -230,18 +367,19 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
       /* keep raw body */
     }
     const surfaced = redact(message.slice(0, 2000) || `Upstream returned ${response.status}`);
-    debug("error", `status=${response.status}`, `body=${surfaced}`);
+    debugWhen(dbg, "error", `status=${response.status}`, `body=${surfaced}`);
     return new Error(surfaced);
   }
 
   async doStream(options: LanguageModelV3CallOptions): Promise<LanguageModelV3StreamResult> {
+    const dbg = resolveDebug(options);
     const body = transform(options, this.modelId);
-    debug("doStream", `model=${this.modelId}`, `bodyBytes=${Buffer.byteLength(body, "utf8")}`);
-    const response = await this.fetchWithRetry(body, options);
-    if (!response.ok || !response.body) throw await this.errorFrom(response);
-    debug("doStream", "ok", `status=${response.status}`);
+    debugWhen(dbg, "doStream", `model=${this.modelId}`, `bodyBytes=${Buffer.byteLength(body, "utf8")}`);
+    const { response, prefix, events } = await this.fetchWithRetry(body, options, dbg);
+    if (!response.ok || !response.body) throw await this.errorFrom(response, dbg);
+    debugWhen(dbg, "doStream", "ok", `status=${response.status}`);
 
-    const stream = toReadableStream(this.streamParts(response.body));
+    const stream = toReadableStream(this.streamParts(prefix, events, dbg));
     return {
       stream,
       request: { body: JSON.parse(body) as unknown },
@@ -250,6 +388,7 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
   }
 
   async doGenerate(options: LanguageModelV3CallOptions): Promise<LanguageModelV3GenerateResult> {
+    const dbg = resolveDebug(options);
     const { stream, request, response } = await this.doStream(options);
 
     let text = "";
@@ -257,6 +396,7 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
     const toolCalls: Array<Extract<LanguageModelV3Content, { type: "tool-call" }>> = [];
     let usage: LanguageModelV3Usage = zeroUsage();
     let finishReason: LanguageModelV3FinishReason = { unified: "other", raw: undefined };
+    let providerMetadata: SharedV3ProviderMetadata | undefined;
     let warnings: SharedV3Warning[] = [];
 
     const reader = stream.getReader();
@@ -280,6 +420,7 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
           case "finish":
             usage = value.usage;
             finishReason = value.finishReason;
+            providerMetadata = value.providerMetadata;
             break;
           case "error":
             throw value.error instanceof Error ? value.error : new Error(String(value.error));
@@ -296,7 +437,8 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
     if (text) content.push({ type: "text", text });
     content.push(...toolCalls);
 
-    debug(
+    debugWhen(
+      dbg,
       "doGenerate",
       "done",
       `model=${this.modelId}`,
@@ -305,12 +447,17 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
       `toolCalls=${toolCalls.length}`,
       `finish=${finishReason.unified ?? "?"}`,
       `usage=${JSON.stringify(usage)}`,
+      `providerMetadata=${JSON.stringify(providerMetadata ?? {})}`,
     );
 
-    return { content, finishReason, usage, warnings, request, response };
+    return { content, finishReason, usage, warnings, request, response, ...(providerMetadata ? { providerMetadata } : {}) };
   }
 
-  private async *streamParts(body: ReadableStream<Uint8Array>): AsyncGenerator<LanguageModelV3StreamPart> {
+  private async *streamParts(
+    prefix: Record<string, any>[],
+    events: AsyncGenerator<Record<string, any>>,
+    dbg: boolean,
+  ): AsyncGenerator<LanguageModelV3StreamPart> {
     yield { type: "stream-start", warnings: [] };
 
     const textId = "text-0";
@@ -321,11 +468,13 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
     let hadToolCalls = false;
     let usage: LanguageModelV3Usage | undefined;
     let finishReason: LanguageModelV3FinishReason | undefined;
+    let cost: number | undefined;
+    let marketCost: number | undefined;
     let errored = false;
 
-    for await (const evt of iterateEvents(body)) {
+    for await (const evt of replayEvents(prefix, events)) {
       const payload = JSON.stringify(evt);
-      debug("stream", `event=${evt.type}`, `payload=${payload && payload.length > 4096 ? payload.slice(0, 4096) + "…" : payload}`);
+      debugWhen(dbg, "stream", `event=${evt.type}`, `payload=${payload && payload.length > 4096 ? payload.slice(0, 4096) + "…" : payload}`);
       switch (evt.type) {
         case "text-start":
           if (!textOpen) {
@@ -395,6 +544,19 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
           usage = usageFromFinish(evt as FinishEvent);
           finishReason = finishReasonFrom(evt as FinishEvent, hadToolCalls);
           break;
+        case "finish-step": {
+          // Upstream dollar cost rides here (usage.raw.cost), never on `finish`.
+          const found = costFromFinishStep(evt as unknown as FinishStepEvent);
+          if (found.cost !== undefined) cost = found.cost;
+          if (found.marketCost !== undefined) marketCost = found.marketCost;
+          break;
+        }
+        case "provider-metadata": {
+          const found = costFromProviderMetadata(evt as unknown as ProviderMetadataEvent);
+          if (cost === undefined && found.cost !== undefined) cost = found.cost;
+          if (marketCost === undefined && found.marketCost !== undefined) marketCost = found.marketCost;
+          break;
+        }
         case "error": {
           errored = true;
           const err = evt.error;
@@ -416,16 +578,27 @@ class CommandCodeLanguageModel implements LanguageModelV3 {
     }
 
     if (errored) {
-      debug("stream", "terminal", "errored");
+      debugWhen(dbg, "stream", "terminal", "errored");
       return;
     }
     if (textOpen) yield { type: "text-end", id: textId };
     if (reasoningOpen) yield { type: "reasoning-end", id: reasoningId };
-    debug("stream", "terminal", `finish=${finishReason ? (finishReason.unified ?? "?") : "synthesized"}`, `usage=${JSON.stringify(usage ?? zeroUsage())}`);
+    const finalUsage = usage ?? zeroUsage();
+    // A reported cost of 0 is meaningful (flat-fee routed request); only
+    // `undefined` means "upstream sent no cost".
+    const reported = {
+      ...(cost !== undefined ? { cost } : {}),
+      ...(marketCost !== undefined ? { marketCost } : {}),
+    };
+    const hasCost = cost !== undefined || marketCost !== undefined;
+    if (hasCost) finalUsage.raw = { ...(finalUsage.raw ?? {}), ...reported };
+    const finishMetadata: SharedV3ProviderMetadata | undefined = hasCost ? { commandcode: reported } : undefined;
+    debugWhen(dbg, "stream", "terminal", `finish=${finishReason ? (finishReason.unified ?? "?") : "synthesized"}`, `usage=${JSON.stringify(finalUsage)}`, `providerMetadata=${JSON.stringify(finishMetadata ?? {})}`);
     yield {
       type: "finish",
-      usage: usage ?? zeroUsage(),
+      usage: finalUsage,
       finishReason: finishReason ?? { unified: hadToolCalls ? "tool-calls" : "stop", raw: undefined },
+      ...(finishMetadata ? { providerMetadata: finishMetadata } : {}),
     };
   }
 }

src/quota.ts

@@ -0,0 +1,431 @@
+// Live Command Code quota from the alpha billing API. Shared by the TUI plugin
+// (loaded from source) and scripts/quota.mjs. Endpoints, validated against
+// api.commandcode.ai:
+//
+//   GET /alpha/whoami
+//   GET /alpha/billing/credits?orgId=
+//   GET /alpha/billing/subscriptions?orgId=
+//   GET /alpha/usage/summary?orgId=
+//
+// These are the same endpoints the `cmd` CLI `/usage` command reads. This module uses
+// global fetch only, so the built provider stays runtime-dependency-free.
+
+import { DEFAULT_BASE_URL } from "./constants.js";
+import { redact } from "./redact.js";
+
+export const QUOTA_TIMEOUT_MS = 15_000;
+
+export type QuotaWindowId = "fiveHour" | "weekly" | "monthly";
+
+export type QuotaWindow = {
+  id: QuotaWindowId;
+  label: string;
+  used: number;
+  cap: number;
+  resetAtMs: number | null;
+  exceeded: boolean;
+};
+
+export type QuotaCredits = {
+  monthly: number;
+  purchased: number;
+  free: number;
+  remaining: number;
+};
+
+export type QuotaAccount = {
+  login: string;
+  keyName?: string;
+  orgId: string | null;
+};
+
+export type QuotaPlan = {
+  id: string;
+  status: string;
+  currentPeriodStart?: string;
+  currentPeriodEnd?: string;
+};
+
+export type QuotaSummary = {
+  totalCost: number;
+  totalCount: number;
+  totalTokens?: number;
+  periodBasis?: string;
+};
+
+export type Quota = {
+  account: QuotaAccount;
+  plan: QuotaPlan | null;
+  credits: QuotaCredits | null;
+  summary: QuotaSummary | null;
+  windows: QuotaWindow[];
+  unavailable: QuotaWindowId[];
+  fetchedAt: number;
+};
+
+export type QuotaErrorKind = "config" | "http" | "network" | "timeout";
+
+export type QuotaResult =
+  | { ok: true; quota: Quota }
+  | { ok: false; error: { kind: QuotaErrorKind; message: string } };
+
+export type QuotaFetchOptions = {
+  apiKey: string;
+  baseURL?: string;
+  orgId?: string;
+  timeoutMs?: number;
+  fetchImpl?: typeof fetch;
+};
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === "object" && value !== null && !Array.isArray(value);
+}
+
+function numberValue(value: unknown): number | undefined {
+  return typeof value === "number" && Number.isFinite(value) ? value : undefined;
+}
+
+function stringValue(value: unknown): string | undefined {
+  return typeof value === "string" && value.length > 0 ? value : undefined;
+}
+
+/** Epoch ms. Upstream has shipped both seconds and ms; >= 1e12 is treated as ms. */
+function resetAtMs(value: unknown): number | null {
+  let n: number | undefined;
+  if (typeof value === "number" && Number.isFinite(value)) n = value;
+  else if (typeof value === "string" && value.trim().length > 0) {
+    const text = value.trim();
+    n = /^\d+$/.test(text) ? Number(text) : Date.parse(text);
+  }
+  if (n === undefined || !Number.isFinite(n) || n < 0) return null;
+  return n >= 1e12 ? Math.round(n) : Math.round(n * 1000);
+}
+
+function parseWindow(id: QuotaWindowId, label: string, value: unknown): QuotaWindow | null {
+  if (!isRecord(value)) return null;
+  const used = numberValue(value.used);
+  const cap = numberValue(value.cap);
+  if (used === undefined || cap === undefined || (used === 0 && cap === 0)) return null;
+  return {
+    id,
+    label,
+    used,
+    cap,
+    resetAtMs: resetAtMs(value.resetAt),
+    exceeded: value.exceeded === true || (cap > 0 && used >= cap),
+  };
+}
+
+function parseCredits(value: unknown): { credits: QuotaCredits; windows: QuotaWindow[] } | null {
+  if (!isRecord(value) || !isRecord(value.credits)) return null;
+  const c = value.credits;
+  const monthly = numberValue(c.monthlyCredits) ?? 0;
+  const purchased = numberValue(c.purchasedCredits) ?? 0;
+  const free = numberValue(c.freeCredits) ?? 0;
+  const limits = isRecord(value.windowLimits) ? value.windowLimits : {};
+  const windows: QuotaWindow[] = [];
+  const fiveHour = parseWindow("fiveHour", "5-hour", limits.fiveHour);
+  const weekly = parseWindow("weekly", "Weekly", limits.weekly);
+  if (fiveHour) windows.push(fiveHour);
+  if (weekly) windows.push(weekly);
+  return { credits: { monthly, purchased, free, remaining: monthly + purchased + free }, windows };
+}
+
+function parseAccount(value: unknown): QuotaAccount | null {
+  if (!isRecord(value)) return null;
+  const org = isRecord(value.org) ? value.org : undefined;
+  const user = isRecord(value.user) ? value.user : undefined;
+  const login =
+    (user ? (stringValue(user.userName) ?? stringValue(user.name)) : undefined) ??
+    (org ? stringValue(org.login) : undefined);
+  if (!login) return null;
+  const keyName = user ? (stringValue(user.keyName) ?? stringValue(user.displayName)) : undefined;
+  return { login, orgId: org ? (stringValue(org.id) ?? null) : null, ...(keyName ? { keyName } : {}) };
+}
+
+function parsePlan(value: unknown): QuotaPlan | null {
+  if (!isRecord(value) || !isRecord(value.data)) return null;
+  const data = value.data;
+  const id = stringValue(data.planId);
+  const status = stringValue(data.status);
+  if (!id && !status) return null;
+  const start = stringValue(data.currentPeriodStart);
+  const end = stringValue(data.currentPeriodEnd);
+  return {
+    id: id ?? "unknown",
+    status: status ?? "unknown",
+    ...(start ? { currentPeriodStart: start } : {}),
+    ...(end ? { currentPeriodEnd: end } : {}),
+  };
+}
+
+function parseSummary(value: unknown): QuotaSummary | null {
+  if (!isRecord(value)) return null;
+  const totalCost = numberValue(value.totalCost);
+  const totalCount = numberValue(value.totalCount);
+  if (totalCost === undefined || totalCount === undefined) return null;
+  const totalTokens = numberValue(value.totalTokens);
+  const periodBasis = stringValue(value.periodBasis);
+  return {
+    totalCost,
+    totalCount,
+    ...(totalTokens === undefined ? {} : { totalTokens }),
+    ...(periodBasis === undefined ? {} : { periodBasis }),
+  };
+}
+
+function periodEndMs(plan: QuotaPlan | null): number | null {
+  if (!plan?.currentPeriodEnd) return null;
+  return resetAtMs(plan.currentPeriodEnd);
+}
+
+class QuotaHttpError extends Error {
+  readonly status: number;
+  readonly body: string;
+  constructor(status: number, body: string) {
+    super(`HTTP ${status}`);
+    this.name = "QuotaHttpError";
+    this.status = status;
+    this.body = body;
+  }
+}
+
+class QuotaTimeoutError extends Error {
+  constructor() {
+    super("Command Code quota request timed out");
+    this.name = "QuotaTimeoutError";
+  }
+}
+
+function isAuthError(error: unknown): boolean {
+  return error instanceof QuotaHttpError && (error.status === 401 || error.status === 403);
+}
+
+export async function fetchQuota(options: QuotaFetchOptions): Promise<QuotaResult> {
+  if (!options.apiKey) {
+    return { ok: false, error: { kind: "config", message: "No Command Code API key found" } };
+  }
+
+  const baseURL = (options.baseURL ?? DEFAULT_BASE_URL).replace(/\/+$/, "");
+  const fetchImpl = options.fetchImpl ?? fetch;
+  const timeoutMs = options.timeoutMs ?? QUOTA_TIMEOUT_MS;
+  const controller = new AbortController();
+  const timer = setTimeout(() => controller.abort(), timeoutMs);
+  const headers = {
+    accept: "application/json",
+    authorization: /^Bearer\s/i.test(options.apiKey) ? options.apiKey : `Bearer ${options.apiKey}`,
+  };
+
+  const request = async (path: string, params?: Record<string, string | undefined>): Promise<unknown> => {
+    if (controller.signal.aborted) throw new QuotaTimeoutError();
+    const search = new URLSearchParams();
+    for (const [key, value] of Object.entries(params ?? {})) if (value) search.set(key, value);
+    const query = search.toString();
+    let response: Response;
+    try {
+      response = await fetchImpl(`${baseURL}${path}${query ? `?${query}` : ""}`, {
+        method: "GET",
+        headers,
+        signal: controller.signal,
+      });
+    } catch (error) {
+      if (controller.signal.aborted) throw new QuotaTimeoutError();
+      throw error;
+    }
+    if (!response.ok) throw new QuotaHttpError(response.status, await response.text().catch(() => ""));
+    return (await response.json()) as unknown;
+  };
+
+  try {
+    const whoamiRaw = await request("/alpha/whoami");
+    const account = parseAccount(whoamiRaw);
+    if (!account) {
+      return {
+        ok: false,
+        error: { kind: "http", message: "Command Code returned an unrecognized account response" },
+      };
+    }
+
+    const orgId = options.orgId ?? account.orgId ?? undefined;
+    const unavailable: QuotaWindowId[] = [];
+
+    let credits: QuotaCredits | null = null;
+    let windows: QuotaWindow[] = [];
+    try {
+      const parsed = parseCredits(await request("/alpha/billing/credits", { orgId }));
+      if (parsed) {
+        credits = parsed.credits;
+        windows = parsed.windows;
+      }
+    } catch (error) {
+      if (isAuthError(error) || error instanceof QuotaTimeoutError) {
+        return { ok: false, error: quotaError(error) };
+      }
+    }
+
+    let plan: QuotaPlan | null = null;
+    try {
+      plan = parsePlan(await request("/alpha/billing/subscriptions", { orgId }));
+    } catch (error) {
+      if (isAuthError(error) || error instanceof QuotaTimeoutError) {
+        return { ok: false, error: quotaError(error) };
+      }
+    }
+
+    let summary: QuotaSummary | null = null;
+    try {
+      summary = parseSummary(await request("/alpha/usage/summary", { orgId, since: plan?.currentPeriodStart }));
+    } catch (error) {
+      if (isAuthError(error) || error instanceof QuotaTimeoutError) {
+        return { ok: false, error: quotaError(error) };
+      }
+    }
+
+    if (credits) {
+      const spent = summary?.totalCost;
+      if (spent !== undefined) {
+        windows.push({
+          id: "monthly",
+          label: "Monthly",
+          used: spent,
+          cap: spent + credits.remaining,
+          resetAtMs: periodEndMs(plan),
+          exceeded: credits.remaining <= 0,
+        });
+      } else {
+        unavailable.push("monthly");
+      }
+    } else {
+      unavailable.push("monthly");
+    }
+
+    if (!credits && !summary) {
+      return {
+        ok: false,
+        error: {
+          kind: "http",
+          message: "Command Code returned no recognized usage data for the account",
+        },
+      };
+    }
+
+    return {
+      ok: true,
+      quota: {
+        account,
+        plan,
+        credits,
+        summary,
+        windows,
+        unavailable,
+        fetchedAt: Date.now(),
+      },
+    };
+  } catch (error) {
+    return { ok: false, error: quotaError(error) };
+  } finally {
+    clearTimeout(timer);
+  }
+}
+
+function quotaError(error: unknown): { kind: QuotaErrorKind; message: string } {
+  if (error instanceof QuotaTimeoutError) return { kind: "timeout", message: error.message };
+  if (error instanceof QuotaHttpError) {
+    const detail = error.body.trim().slice(0, 200);
+    const hint = error.status === 401 || error.status === 403 ? " (check the API key)" : "";
+    return {
+      kind: "http",
+      message: redact(`Command Code quota request failed (${error.status})${hint}: ${detail}`),
+    };
+  }
+  return {
+    kind: "network",
+    message: redact(`Command Code quota request failed: ${error instanceof Error ? error.message : String(error)}`),
+  };
+}
+
+export function percent(used: number, cap: number): number {
+  if (!(cap > 0)) return 0;
+  return Math.round((used / cap) * 100);
+}
+
+export function quotaBar(used: number, cap: number, width = 10): string {
+  const fraction = cap > 0 ? Math.min(1, Math.max(0, used / cap)) : 0;
+  const filled = Math.round(fraction * width);
+  return `${"█".repeat(filled)}${"░".repeat(width - filled)}`;
+}
+
+/** Compact countdown, e.g. `3h 12m` / `2d 4h`. Empty when there is no reset. */
+export function formatReset(resetAtMs: number | null, now: number = Date.now()): string {
+  if (resetAtMs === null) return "";
+  const diff = resetAtMs - now;
+  if (diff <= 0) return "soon";
+  const minutes = Math.ceil(diff / 60_000);
+  if (minutes < 60) return `${minutes}m`;
+  const hours = Math.floor(minutes / 60);
+  const mins = minutes % 60;
+  if (hours < 24) return mins > 0 ? `${hours}h ${mins}m` : `${hours}h`;
+  const days = Math.floor(hours / 24);
+  const rem = hours % 24;
+  return rem > 0 ? `${days}d ${rem}h` : `${days}d`;
+}
+
+function formatTokens(tokens: number): string {
+  if (tokens >= 1_000_000_000) return `${(tokens / 1_000_000_000).toFixed(1)}B`;
+  if (tokens >= 1_000_000) return `${(tokens / 1_000_000).toFixed(1)}M`;
+  if (tokens >= 1_000) return `${(tokens / 1_000).toFixed(1)}k`;
+  return String(tokens);
+}
+
+function formatWindowLine(w: QuotaWindow, now: number): string {
+  const pct = Math.min(percent(w.used, w.cap), 999);
+  const reset = formatReset(w.resetAtMs, now);
+  const amount =
+    w.id === "monthly"
+      ? `$${w.used.toFixed(2)} / $${w.cap.toFixed(2)}`
+      : `${w.used.toFixed(2)} / ${w.cap.toFixed(2)} credits`;
+  return `${w.label.padEnd(8)} ${quotaBar(w.used, w.cap)} ${String(pct).padStart(3)}%  ${amount}${
+    reset ? `  (resets in ${reset})` : ""
+  }`;
+}
+
+function planLine(plan: QuotaPlan, now: number): string {
+  const name = plan.id.replace(/[_-]+/g, " ").replace(/\b\w/g, (c) => c.toUpperCase());
+  const status = plan.status ? ` · ${plan.status}` : "";
+  const end = periodEndMs(plan);
+  if (end === null) return `Plan: ${name}${status}`;
+  const remaining = end - now;
+  const days = Math.ceil(remaining / 86_400_000);
+  const date = new Date(end).toISOString().slice(0, 10);
+  const when = days > 0 ? `renews ${date} (${days}d)` : days === 0 ? `renews ${date} (today)` : `renewed ${date}`;
+  return `Plan: ${name}${status} · ${when}`;
+}
+
+/** Human-readable multi-line quota summary for the `/cc-usage` command and CLI. */
+export function formatQuota(quota: Quota, now: number = Date.now()): string {
+  const lines: string[] = [];
+  lines.push(quota.account.keyName ?? quota.account.login);
+  if (quota.plan) lines.push(planLine(quota.plan, now));
+
+  if (quota.windows.length > 0) {
+    lines.push("");
+    for (const w of quota.windows) lines.push(formatWindowLine(w, now));
+  }
+
+  if (quota.credits) {
+    const parts = [`monthly $${quota.credits.monthly.toFixed(2)}`, `purchased $${quota.credits.purchased.toFixed(2)}`];
+    if (quota.credits.free > 0) parts.push(`free $${quota.credits.free.toFixed(2)}`);
+    lines.push("", `Credits: ${parts.join(" / ")}`);
+  }
+
+  if (quota.summary) {
+    const period = quota.plan?.currentPeriodStart ? "this billing period" : "total";
+    const tokens = quota.summary.totalTokens === undefined ? "" : ` · ${formatTokens(quota.summary.totalTokens)} tokens`;
+    lines.push(
+      `Usage: ${quota.summary.totalCount.toLocaleString("en-US")} requests · $${quota.summary.totalCost.toFixed(2)} (${period})${tokens}`,
+    );
+  }
+
+  if (quota.unavailable.length > 0) lines.push("", `Unavailable: ${quota.unavailable.join(", ")}`);
+  return lines.join("\n");
+}

src/toggles.ts

@@ -0,0 +1,55 @@
+// Runtime toggle state shared between the provider (server side) and the slash-command
+// TUI plugin (client side) through a small JSON flag file. Node built-ins only so the
+// built provider stays dependency-free; never throws on a missing or malformed file.
+
+import { mkdirSync, readFileSync, renameSync, writeFileSync } from "node:fs";
+import { homedir } from "node:os";
+import { dirname, join } from "node:path";
+
+export type Toggles = { zdr?: boolean; debug?: boolean };
+
+export type ToggleName = keyof Toggles;
+
+export function togglesPath(): string {
+  const override = process.env["COMMANDCODE_TOGGLES_FILE"];
+  if (override) return override;
+  const configHome = process.env["XDG_CONFIG_HOME"] ?? join(homedir(), ".config");
+  return join(configHome, "opencode", "commandcode-toggles.json");
+}
+
+export function readToggles(): Toggles {
+  try {
+    const parsed = JSON.parse(readFileSync(togglesPath(), "utf8")) as Record<string, unknown>;
+    if (!parsed || typeof parsed !== "object") return {};
+    const out: Toggles = {};
+    if (typeof parsed["zdr"] === "boolean") out.zdr = parsed["zdr"];
+    if (typeof parsed["debug"] === "boolean") out.debug = parsed["debug"];
+    return out;
+  } catch {
+    return {};
+  }
+}
+
+export function writeToggles(patch: Toggles): Toggles {
+  const merged = { ...readToggles(), ...patch };
+  const path = togglesPath();
+  mkdirSync(dirname(path), { recursive: true });
+  const tmp = `${path}.${process.pid}.tmp`;
+  writeFileSync(tmp, `${JSON.stringify(merged, null, 2)}\n`, "utf8");
+  renameSync(tmp, path);
+  return merged;
+}
+
+export function toggle(name: ToggleName): Toggles {
+  const current = readToggles();
+  return writeToggles({ [name]: current[name] !== true });
+}
+
+/** Toggle state from the shared flag file; an absent key means off. */
+export function zdrEnabled(): boolean {
+  return readToggles().zdr === true;
+}
+
+export function debugEnabled(): boolean {
+  return readToggles().debug === true;
+}

src/tui-shims.d.ts

@@ -0,0 +1,19 @@
+// Ambient shims for the symbols the TUI plugin imports from opencode's bundled runtime.
+// opencode resolves `solid-js` and `@opentui/solid` to its own internal modules when it
+// loads the `.tsx` plugin, so this package declares just enough surface for `tsc` without
+// taking a dependency. The JSX namespace is intentionally permissive: only a handful of
+// intrinsic elements are used and their props come from opencode's theme at runtime.
+
+declare module "solid-js" {
+  export function createSignal<Value>(value: Value): [get: () => Value, set: (value: Value) => void];
+}
+
+declare namespace JSX {
+  type Element = unknown;
+  interface ElementChildrenAttribute {
+    children: {};
+  }
+  interface IntrinsicElements {
+    [name: string]: any;
+  }
+}

src/tui.tsx

@@ -0,0 +1,410 @@
+// opencode TUI plugin exposing /cc-zdr, /cc-debug, /cc-status and /cc-usage, plus a sidebar
+// panel showing the current toggle state and live CommandCode quota. It flips the shared toggle
+// file (see ./toggles.ts) that the provider reads on every request, so a change takes effect
+// without restarting opencode, and reads the alpha billing endpoints (see ./quota.ts) with the
+// provider's API key for the 5-hour, weekly and monthly meters.
+//
+// opencode loads this module's default export (`{ id, tui }`) from a `tui.json` plugin
+// entry. It is loaded from source (not dist) so opencode's Bun/Solid transform compiles
+// the JSX and maps `solid-js` / `@opentui/solid` to its internal modules; that keeps this
+// package free of runtime and build dependencies. The ambient types in ./tui-shims.d.ts
+// back the JSX and `createSignal` references for `tsc -p tsconfig.tui.json`.
+//
+// Note: in this opencode version slash commands and the Ctrl+P palette read the same
+// `namespace: "palette"` command list, so these entries appear in both; `hidden: true`
+// would remove them from both.
+
+import { appendFileSync } from "node:fs";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { createSignal } from "solid-js";
+
+import { fetchQuota, formatQuota, formatReset, percent, quotaBar, type QuotaResult, type QuotaWindow } from "./quota.js";
+import { redact } from "./redact.js";
+import { debugEnabled, readToggles, toggle, type ToggleName, type Toggles } from "./toggles.js";
+
+type ToastVariant = "info" | "success" | "warning" | "error";
+
+type TuiCommand = {
+  name: string;
+  title?: string;
+  desc?: string;
+  category?: string;
+  namespace?: string;
+  slashName?: string;
+  slashAliases?: string[];
+  suggested?: boolean;
+  hidden?: boolean;
+  enabled?: boolean | (() => boolean);
+  run: () => void | Promise<void>;
+};
+
+type TuiTheme = {
+  current: { text: unknown; textMuted: unknown; success: unknown; warning: unknown; error: unknown };
+};
+
+type TuiSlotHandler = (ctx: { theme: TuiTheme }, props: { session_id: string }) => unknown;
+
+type TuiSlotPlugin = { order?: number; slots: Record<string, TuiSlotHandler> };
+
+type TuiProvider = { id?: string; key?: string; options?: Record<string, unknown> };
+
+type TuiProviderConfig = { options?: Record<string, unknown> };
+
+type TuiState = {
+  config?: { provider?: Record<string, TuiProviderConfig | undefined> };
+  provider?: ReadonlyArray<TuiProvider>;
+};
+
+type TuiApi = {
+  keymap: { registerLayer(layer: { commands?: readonly TuiCommand[] }): () => void };
+  ui: { toast(input: { title?: string; message: string; variant?: ToastVariant; duration?: number }): void };
+  slots: { register(plugin: TuiSlotPlugin): string };
+  event: { on(type: string, handler: (event?: { type?: string }) => void): () => void };
+  theme: TuiTheme;
+  state: TuiState;
+  lifecycle: { onDispose(fn: () => void): () => void };
+};
+
+const ID = "commandcode-toggles";
+const CATEGORY = "CommandCode";
+const SIDEBAR_ORDER = 90;
+const DEFAULT_QUOTA_INTERVAL_MS = 300_000;
+const MIN_REFRESH_INTERVAL_MS = 120_000;
+// Extra TUI-bus signals that mark the end of a turn. `session.idle` is a server
+// plugin event and may never fire here; these keep the panel fresh regardless.
+// All event-driven refreshes are floored to MIN_REFRESH_INTERVAL_MS (hard skip).
+const QUOTA_TRIGGER_EVENTS = ["session.idle", "session.status", "session.updated", "message.updated"] as const;
+
+function safeEnv(name: string): string | undefined {
+  try {
+    return typeof process === "undefined" ? undefined : process.env?.[name];
+  } catch {
+    return undefined;
+  }
+}
+
+/** Debug-only trace to the shared debug file. Silent unless debug is on; never logs the key. */
+function trace(...args: unknown[]): void {
+  if (!debugEnabled()) return;
+  const file = join(tmpdir(), "commandcode-debug.log");
+  const parts = args.map((value) => (typeof value === "string" ? value : safeJson(value)));
+  const line = `[commandcode] ${new Date().toISOString()} [tui-quota] ${parts.join(" ")}`;
+  try {
+    appendFileSync(file, redact(line) + "\n", "utf8");
+  } catch {
+    // Never let tracing break the panel.
+  }
+}
+
+function safeJson(value: unknown): string {
+  try {
+    return JSON.stringify(value);
+  } catch {
+    return String(value);
+  }
+}
+
+function state(value: boolean | undefined): string {
+  return value === true ? "on" : "off";
+}
+
+function summary(toggles: Toggles = readToggles()): string {
+  return `zdr=${state(toggles.zdr)}, debug=${state(toggles.debug)}`;
+}
+
+/** Expands an `{env:NAME}` config reference; other templates are treated as unset. */
+function resolveRef(value: unknown): string | undefined {
+  if (typeof value !== "string" || value.length === 0) return undefined;
+  const trimmed = value.trim();
+  const template = /^\{env:([A-Za-z_][A-Za-z0-9_]*)\}$/.exec(trimmed);
+  if (template) return template[1] ? safeEnv(template[1]) : undefined;
+  if (/^\{.*\}$/.test(trimmed)) return undefined;
+  return trimmed;
+}
+
+function commandCodeOptions(api: TuiApi): Record<string, unknown> | undefined {
+  try {
+    return api.state?.config?.provider?.["commandcode"]?.options;
+  } catch {
+    return undefined;
+  }
+}
+
+function providersOf(api: TuiApi): TuiProvider[] {
+  try {
+    const list = api.state?.provider;
+    if (Array.isArray(list)) return list as TuiProvider[];
+    // Tolerate a non-array provider state shape instead of throwing.
+    if (list && typeof list === "object") return Object.values(list) as TuiProvider[];
+  } catch {
+    /* fall through */
+  }
+  return [];
+}
+
+function resolveApiKeySource(api: TuiApi): { key?: string; source: string } {
+  const provider = providersOf(api).find((entry) => entry?.id === "commandcode");
+  const options = commandCodeOptions(api);
+  const headers = options?.["headers"] as Record<string, unknown> | undefined;
+  const named: Array<[string, unknown]> = [
+    ["provider.key", provider?.key],
+    ["provider.options.apiKey", provider?.options?.["apiKey"]],
+    ["headers.Authorization", headers?.["Authorization"] ?? headers?.["authorization"]],
+    ["options.apiKey", options?.["apiKey"]],
+    ["COMMANDCODE_API_KEY", safeEnv("COMMANDCODE_API_KEY")],
+  ];
+  for (const [source, candidate] of named) {
+    const value = resolveRef(candidate);
+    if (value) return { key: value.replace(/^Bearer\s+/i, ""), source };
+  }
+  return { source: "none" };
+}
+
+function resolveBaseURL(api: TuiApi): string | undefined {
+  return resolveRef(commandCodeOptions(api)?.["baseURL"]);
+}
+
+function quotaIntervalMs(): number {
+  const raw = Number(safeEnv("COMMANDCODE_QUOTA_INTERVAL_MS"));
+  return Number.isFinite(raw) && raw > 0 ? raw : DEFAULT_QUOTA_INTERVAL_MS;
+}
+
+function formatClock(ms: number): string {
+  if (!(ms > 0)) return "--:--:--";
+  return new Date(ms).toTimeString().slice(0, 8);
+}
+
+function shortWindow(w: QuotaWindow, nowMs: number): string {
+  const label = w.id === "fiveHour" ? "5h" : w.id === "weekly" ? "7d" : "mo";
+  const reset = formatReset(w.resetAtMs, nowMs);
+  const tail = reset ? ` · ${reset}` : "";
+  const amount = w.id === "monthly" ? `$${w.used.toFixed(2)}/$${w.cap.toFixed(2)}` : `${percent(w.used, w.cap)}%`;
+  return `${label} ${quotaBar(w.used, w.cap, 8)} ${amount}${tail}`;
+}
+
+export const tui = async (api: TuiApi): Promise<void> => {
+  const [toggles, setToggles] = createSignal(readToggles());
+  const [quota, setQuota] = createSignal<QuotaResult | null>(null);
+  const [now, setNow] = createSignal(Date.now());
+  const [updatedAt, setUpdatedAt] = createSignal(0);
+
+  let inflight = false;
+  let pending = false;
+  let lastRefreshAt = 0;
+  let debounceTimer: ReturnType<typeof setTimeout> | undefined;
+  let lastSessionId: string | undefined;
+
+  const refresh = (): void => setToggles(readToggles());
+
+  const refreshQuota = async (): Promise<void> => {
+    if (inflight) {
+      pending = true;
+      trace("coalesced", "inflight");
+      return;
+    }
+    const since = Date.now() - lastRefreshAt;
+    if (lastRefreshAt > 0 && since < MIN_REFRESH_INTERVAL_MS) {
+      trace("rate-limited", `remainingMs=${MIN_REFRESH_INTERVAL_MS - since}`);
+      return;
+    }
+    inflight = true;
+    try {
+      const { key: apiKey, source } = resolveApiKeySource(api);
+      if (!apiKey) {
+        setQuota({ ok: false, error: { kind: "config", message: "No API key found" } });
+        setNow(Date.now());
+        trace("no-key", `source=${source}`);
+        return;
+      }
+      trace("fetch-start", `source=${source}`);
+      const result = await fetchQuota({ apiKey, baseURL: resolveBaseURL(api) });
+      setQuota(result);
+      if (result.ok) {
+        trace(
+          "fetch-ok",
+          `windows=${result.quota.windows.map((w) => `${w.id}:${w.used}/${w.cap}`).join(",")}`,
+        );
+      } else {
+        trace("fetch-error", `kind=${result.error.kind}`);
+      }
+    } catch (error) {
+      // Key resolution, state access, and fetch can all throw (e.g. unsynced
+      // api.state); surface it instead of wedging the panel on `loading…`.
+      const message = redact(error instanceof Error ? error.message : String(error));
+      setQuota({ ok: false, error: { kind: "network", message } });
+      trace("fetch-throw", message);
+    } finally {
+      inflight = false;
+      lastRefreshAt = Date.now();
+      setNow(Date.now());
+      setUpdatedAt(Date.now());
+      if (pending) {
+        pending = false;
+        trace("flush-pending");
+        void refreshQuota();
+      }
+    }
+  };
+
+  const scheduleQuotaRefresh = (reason: string, delayMs = 800): void => {
+    trace("scheduled", `reason=${reason}`, `delayMs=${delayMs}`);
+    if (debounceTimer !== undefined) clearTimeout(debounceTimer);
+    debounceTimer = setTimeout(() => {
+      debounceTimer = undefined;
+      void refreshQuota();
+    }, delayMs);
+  };
+
+  const flip = (name: ToggleName): void => {
+    const next = toggle(name);
+    setToggles(next);
+    api.ui.toast({
+      title: CATEGORY,
+      message: `${name} ${state(next[name])} (${summary(next)})`,
+      variant: next[name] === true ? "success" : "info",
+    });
+  };
+
+  function tone(w: QuotaWindow, theme: TuiTheme): unknown {
+    if (w.exceeded || percent(w.used, w.cap) >= 100) return theme.current.error;
+    if (percent(w.used, w.cap) >= 80) return theme.current.warning;
+    return theme.current.success;
+  }
+
+  // Reads quota()/now() inside the JSX so Solid tracks the signals. Hoisting
+  // `quota()` into a `const` above the return would subscribe once under a
+  // non-tracking owner (the slot re-invokes the handler on session change,
+  // which is why /session "fixed" it) and freeze the panel after first paint.
+  const windowsOf = (result: QuotaResult | null): QuotaWindow[] =>
+    result?.ok === true ? result.quota.windows : [];
+
+  const errorOf = (result: QuotaResult | null): string | undefined =>
+    result?.ok === false ? result.error.message : undefined;
+
+  const Status = () => {
+    const theme = api.theme;
+    return (
+      <box flexDirection="column" gap={0}>
+        <text fg={theme.current.text}>
+          <b>CommandCode</b>
+        </text>
+        <text fg={theme.current.textMuted}>
+          zdr:{state(toggles().zdr)} debug:{state(toggles().debug)}
+        </text>
+        {windowsOf(quota()).map((w) => (
+          <text fg={tone(w, theme)}>{shortWindow(w, now())}</text>
+        ))}
+        {errorOf(quota()) !== undefined ? (
+          <text fg={theme.current.error}>quota: {errorOf(quota())}</text>
+        ) : quota() === null ? (
+          <text fg={theme.current.textMuted}>quota: loading…</text>
+        ) : null}
+        <text fg={theme.current.textMuted}>updated @ {formatClock(updatedAt())}</text>
+      </box>
+    );
+  };
+
+  api.slots.register({
+    order: SIDEBAR_ORDER,
+    slots: {
+      sidebar_content(_ctx, props?: { session_id?: string }) {
+        const id = props?.session_id;
+        if (id !== undefined && id !== lastSessionId) {
+          lastSessionId = id;
+          scheduleQuotaRefresh("session-change");
+        }
+        return <Status />;
+      },
+    },
+  });
+
+  api.keymap.registerLayer({
+    commands: [
+      {
+        name: `${ID}.zdr`,
+        title: "CommandCode: toggle ZDR header",
+        desc: "Send the x-cmd-zdr header on requests",
+        category: CATEGORY,
+        namespace: "palette",
+        slashName: "cc-zdr",
+        run: () => flip("zdr"),
+      },
+      {
+        name: `${ID}.debug`,
+        title: "CommandCode: toggle debug tracing",
+        desc: "Write a redacted request and stream trace to a log file",
+        category: CATEGORY,
+        namespace: "palette",
+        slashName: "cc-debug",
+        run: () => flip("debug"),
+      },
+      {
+        name: `${ID}.status`,
+        title: "CommandCode: show toggle status",
+        desc: "Show the current zdr and debug state",
+        category: CATEGORY,
+        namespace: "palette",
+        slashName: "cc-status",
+        run: () => api.ui.toast({ title: CATEGORY, message: summary(), variant: "info" }),
+      },
+      {
+        name: `${ID}.usage`,
+        title: "CommandCode: show usage and quota",
+        desc: "Show the 5-hour, weekly and monthly quota from the CommandCode API",
+        category: CATEGORY,
+        namespace: "palette",
+        slashName: "cc-usage",
+        run: async () => {
+          const { key: apiKey } = resolveApiKeySource(api);
+          if (!apiKey) {
+            api.ui.toast({
+              title: CATEGORY,
+              message: "No API key found. Set provider.commandcode.options.apiKey or COMMANDCODE_API_KEY.",
+              variant: "warning",
+            });
+            return;
+          }
+          let result: QuotaResult;
+          try {
+            result = await fetchQuota({ apiKey, baseURL: resolveBaseURL(api) });
+          } catch (error) {
+            const message = redact(error instanceof Error ? error.message : String(error));
+            setQuota({ ok: false, error: { kind: "network", message } });
+            trace("usage-throw", message);
+            api.ui.toast({ title: CATEGORY, message, variant: "error" });
+            return;
+          }
+          if (!result.ok) {
+            setQuota(result);
+            api.ui.toast({ title: CATEGORY, message: result.error.message, variant: "error" });
+            return;
+          }
+          setQuota(result);
+          setNow(Date.now());
+          setUpdatedAt(Date.now());
+          api.ui.toast({ title: CATEGORY, message: formatQuota(result.quota), variant: "info", duration: 15000 });
+        },
+      },
+    ],
+  });
+
+  void refreshQuota();
+  const toggleTimer = setInterval(refresh, 1500);
+  const clockTimer = setInterval(() => setNow(Date.now()), 30_000);
+  const quotaTimer = setInterval(() => void refreshQuota(), quotaIntervalMs());
+  const unsubs = QUOTA_TRIGGER_EVENTS.map((type) =>
+    api.event.on(type, (event) => scheduleQuotaRefresh(event?.type ?? type)),
+  );
+  api.lifecycle.onDispose(() => {
+    for (const unsub of unsubs) unsub();
+    if (debounceTimer !== undefined) clearTimeout(debounceTimer);
+    clearInterval(toggleTimer);
+    clearInterval(clockTimer);
+    clearInterval(quotaTimer);
+  });
+};
+
+export const id = ID;
+
+export default { id: ID, tui };

src/usage.ts

@@ -6,8 +6,15 @@ export type CommandCodeUsage = {
   outputTokens?: number;
   totalTokens?: number;
   cachedInputTokens?: number;
+  reasoningTokens?: number;
+  cost?: number | string;
+  market_cost?: number | string;
+  marketCost?: number | string;
+  gateway_cost?: number | string;
+  gatewayCost?: number | string;
   inputTokenDetails?: { noCacheTokens?: number; cacheReadTokens?: number };
   outputTokenDetails?: { textTokens?: number; reasoningTokens?: number };
+  raw?: Record<string, unknown>;
 };
 
 export type FinishEvent = {
@@ -16,6 +23,73 @@ export type FinishEvent = {
   totalUsage?: CommandCodeUsage;
 };
 
+export type FinishStepEvent = {
+  usage?: CommandCodeUsage;
+  providerMetadata?: Record<string, unknown>;
+};
+
+export type ProviderMetadataEvent = {
+  providerMetadata?: Record<string, unknown>;
+};
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === "object" && value !== null && !Array.isArray(value);
+}
+
+/** Finite numbers pass through; numeric strings (gateway `cost: "0.0003"`) are coerced. */
+function numberFrom(value: unknown): number | undefined {
+  if (typeof value === "number" && Number.isFinite(value)) return value;
+  if (typeof value === "string" && value.trim().length > 0) {
+    const n = Number(value.trim());
+    if (Number.isFinite(n)) return n;
+  }
+  return undefined;
+}
+
+/**
+ * Upstream dollar cost arrives on `finish-step`, not `finish`:
+ * `usage.raw.{cost,market_cost,gateway_cost}` (numbers). A reported `0` is
+ * meaningful (flat-fee routed request) and must be preserved.
+ */
+export function costFromFinishStep(evt: FinishStepEvent): { cost?: number; marketCost?: number } {
+  const fallback = costFromProviderMetadata(evt);
+  const usage = evt.usage;
+  if (!usage || typeof usage !== "object") return fallback;
+  const raw = isRecord(usage.raw) ? usage.raw : {};
+  const cost = numberFrom(usage.cost ?? raw["cost"]) ?? fallback.cost;
+  const marketCost =
+    numberFrom(
+      usage.market_cost ??
+        usage.marketCost ??
+        raw["market_cost"] ??
+        raw["marketCost"] ??
+        usage.gateway_cost ??
+        usage.gatewayCost ??
+        raw["gateway_cost"] ??
+        raw["gatewayCost"],
+    ) ?? fallback.marketCost;
+  return {
+    ...(cost !== undefined ? { cost } : {}),
+    ...(marketCost !== undefined ? { marketCost } : {}),
+  };
+}
+
+/** Fallback/supplement: `provider-metadata` carries gateway cost as strings. */
+export function costFromProviderMetadata(evt: ProviderMetadataEvent): {
+  cost?: number;
+  marketCost?: number;
+} {
+  const pm = evt.providerMetadata;
+  if (!isRecord(pm)) return {};
+  const gateway = isRecord(pm["gateway"]) ? (pm["gateway"] as Record<string, unknown>) : {};
+  const cost = numberFrom(gateway["cost"]);
+  const marketCost = numberFrom(gateway["marketCost"] ?? gateway["market_cost"]);
+  return {
+    ...(cost !== undefined ? { cost } : {}),
+    ...(marketCost !== undefined ? { marketCost } : {}),
+  };
+}
+
 /** CommandCode `finish` -> LanguageModelV3Usage. */
 export function usageFromFinish(evt: FinishEvent): LanguageModelV3Usage {
   const tu = evt.totalUsage;
@@ -26,10 +100,12 @@ export function usageFromFinish(evt: FinishEvent): LanguageModelV3Usage {
   const details = tu.inputTokenDetails ?? {};
   const outDetails = tu.outputTokenDetails ?? {};
   const cacheRead = tu.cachedInputTokens ?? details.cacheReadTokens;
+  // inputTokens.total is input-only (parity with server.py); totalTokens is input+output.
+  const total = tu.totalTokens ?? input + output;
 
   return {
     inputTokens: {
-      total: tu.totalTokens ?? input + output,
+      total: tu.inputTokens ?? total - output,
       noCache: details.noCacheTokens,
       cacheRead,
       cacheWrite: undefined,

tsconfig.json

@@ -14,5 +14,6 @@
     "noUncheckedIndexedAccess": true,
     "verbatimModuleSyntax": true
   },
-  "include": ["src"]
+  "include": ["src"],
+  "exclude": ["src/tui.tsx", "src/tui-shims.d.ts"]
 }

tsconfig.tui.json

@@ -0,0 +1,18 @@
+{
+  "compilerOptions": {
+    "target": "ES2022",
+    "lib": ["ES2022", "DOM", "DOM.Iterable"],
+    "module": "ESNext",
+    "moduleResolution": "Bundler",
+    "types": ["node"],
+    "strict": true,
+    "noUncheckedIndexedAccess": true,
+    "verbatimModuleSyntax": true,
+    "esModuleInterop": true,
+    "forceConsistentCasingInFileNames": true,
+    "skipLibCheck": true,
+    "jsx": "preserve",
+    "noEmit": true
+  },
+  "include": ["src/tui.tsx", "src/tui-shims.d.ts", "src/quota.ts", "src/toggles.ts", "src/redact.ts", "src/constants.ts"]
+}