Delete @workbench/hub-client (CL-7383)

IMPLEMENTATION.md

@@ -244,7 +244,7 @@ handshake.
 list a workbench host launches with (`workbenchHostInferencePreferences`
 on the chat adapter; also `tenantDefaultModel` on agent-definition
 routes). It keeps credentialed completion-capable offerings
-(`preferCompletionCapable` in `@workbench/hub-client/model-capability`
+(`preferCompletionCapable` in `@workbench/connections/model-capability`
 — embedding names never win) and, when any survivor is
 `origin.direct`, picks from that direct set only. Inherit-only catalogs
 still sort among inherited completion rows.

README.md

@@ -90,7 +90,7 @@ account: anyone who signs up gets a personal bench with no default routines
 deployed, and first-run tells them exactly that. Onboarding walks them
 through picking a provider — Anthropic, OpenAI, Google, OpenRouter, Hugging
 Face, Groq, or another of the curated providers in
-[`packages/hub-client/src/catalog-seed-data.ts`](packages/hub-client/src/catalog-seed-data.ts)
+[`packages/seeding/src/catalog-seed-data.ts`](packages/seeding/src/catalog-seed-data.ts)
 — and pasting their own key (or, for OpenRouter, completing a PKCE OAuth
 connect), which is stored right away — no separate seeding step, no
 docs to read. A wrong key isn't caught up front; it surfaces the first time

apps/hub/package.json

@@ -66,7 +66,6 @@
     "@workbench/connections": "workspace:*",
     "@corbits/hub-api-client": "workspace:*",
     "@corbits/seeding": "workspace:*",
-    "@workbench/hub-client": "workspace:*",
     "@workbench/onboarding": "workspace:*",
     "@workbench/templates": "workspace:*",
     "arktype": "catalog:",

apps/hub/src/config.ts

@@ -44,7 +44,7 @@ import {
   envProviderBaseUrlsFrom,
   envProviderKeysFrom,
 } from "@workbench/onboarding";
-import type { SupportedCredentialProvider } from "@workbench/hub-client";
+import type { SupportedCredentialProvider } from "@workbench/connections/credential-test";
 
 const HTTP_URL = /^https?:\/\/.+$/;
 

apps/hub/src/env-credential-plant.ts

@@ -22,7 +22,7 @@
 // virgin, open-signup database. Falling through to sign-up the way
 // those interactive commands do would let a boot-time retry tick mint
 // the default admin account (and its default password) on its own —
-// so this module authenticates with `@workbench/hub-client`'s
+// so this module authenticates with `@corbits/hub-api-client`'s
 // sign-in-only `signIn`, and an unresolved admin account is treated
 // exactly like an unresolved bench: retried quietly, never a reason to
 // self-provision the account.

apps/hub/src/index.ts

@@ -287,7 +287,7 @@ import {
   createPresenceRoutes,
   type PresenceRoomKey,
 } from "@corbits/presence";
-import { supportedCredentialProviders } from "@workbench/hub-client";
+import { supportedCredentialProviders } from "@workbench/connections/credential-test";
 import { createGitWorkflowPusher } from "@corbits/seeding";
 import { createHubAPI } from "@corbits/hub-api-client";
 import {
@@ -1300,7 +1300,7 @@ export async function createHub(config: HubConfig) {
   // Mounted outside the tenant prefix — the sidecar reaches it as a
   // plain inference endpoint, never through tenant-scoped auth, the
   // same way it reaches a real provider's API. Pinned by the heartbeat
-  // and workbench-digest workflows' seeds (`@corbits/hub-client`), whose
+  // and workbench-digest workflows' seeds (`@corbits/seeding`), whose
   // agents never produce text. `config.baseUrl` (not `localhost`) is
   // what makes the URL usable from a sidecar on another machine.
   app.route("/api/chat/noop-inference", createNoopInferenceRoutes());

apps/sidecar/src/config.ts

@@ -14,7 +14,7 @@ import { parseToolRegistries } from "./tool-materialization";
 /**
  * The shipped default manifest: registers `@corbits/ollama-adapter`'s
  * `createOllamaAdapter` for the `"ollama"` provider key so a seeded
- * Ollama deployment's `quirks.numCtx` (see `@corbits/hub-client`'s
+ * Ollama deployment's `quirks.numCtx` (see `@corbits/seeding`'s
  * seed and `./workflow-substrate-factory/context-budget`) actually
  * reaches Ollama's `options.num_ctx` instead of silently falling back
  * to the built-in adapter's defaults. An operator who sets

apps/sidecar/test/adapter-registry.test.ts

@@ -34,7 +34,7 @@ test("a seeded Ollama model's quirks.numCtx reaches the built request as options
     provider: "ollama",
     model: "gpt-oss:20b",
   };
-  // Shaped exactly like `@corbits/hub-client`'s seed writes onto a
+  // Shaped exactly like `@corbits/seeding`'s seed writes onto a
   // catalog offering's `quirks` column (`quirksForDeployment`).
   const quirks = { default: { numCtx: 32_768 } };
   const adapter = registry.resolve(source, quirks);

apps/web/package.json

@@ -48,7 +48,6 @@
     "@radix-ui/react-slot": "^1.2.3",
     "@tanstack/react-query": "catalog:",
     "@workbench/connections": "workspace:*",
-    "@workbench/hub-client": "workspace:*",
     "@workbench/templates": "workspace:*",
     "arktype": "catalog:",
     "react": "^19.2.0",

apps/web/src/agents-api.ts

@@ -23,7 +23,7 @@ import {
   UnauthenticatedError,
   toAPIQuery,
 } from "@corbits/api-query";
-import { isChatPickerModelName } from "@workbench/hub-client/model-capability";
+import { isChatPickerModelName } from "@workbench/connections/model-capability";
 import { tenantKeys } from "./query-client";
 
 export type AgentDefinition = typeof WorkflowDefinitionResponse.infer;

apps/web/src/myra-workbench.ts

@@ -16,7 +16,7 @@ import { listAgentDefinitions, type AgentDefinition } from "./agents-api";
 
 export const MYRA_WORKBENCH_TITLE = "Myra";
 
-/** The seeded workflow asset backing Myra (`packages/hub-client/src/seed.ts`
+/** The seeded workflow asset backing Myra (`packages/seeding/src/seed.ts`
  * deploys it as `assistant`, stamped with catalog displayName "Myra"). A
  * chat's `definitionId` names this deployed definition's row id, never the
  * asset name itself. */

apps/web/src/onboarding.ts

@@ -6,7 +6,7 @@
 // so a broken provisioning call never leaves the user silently benchless.
 
 import { type } from "arktype";
-import type { SupportedCredentialProvider } from "@workbench/hub-client/credential-test";
+import type { SupportedCredentialProvider } from "@workbench/connections/credential-test";
 
 const ProvisionResult = type({
   kind: "'existing-member' | 'provisioned' | 'needs-onboarding'",

apps/web/src/pages/onboarding-page.tsx

@@ -32,7 +32,7 @@ import { Key, WarningCircle } from "@corbits/icons";
 import { WorkbenchLoadingState } from "@corbits/chat-ui";
 import { useCallback, useEffect, useState } from "react";
 import type { FormEvent } from "react";
-import { OLLAMA_PLACEHOLDER_SECRET } from "@workbench/hub-client/credential-test";
+import { OLLAMA_PLACEHOLDER_SECRET } from "@workbench/connections/credential-test";
 
 import { useNavigate } from "../navigation";
 import {

apps/web/test/onboarding.test.tsx

@@ -13,7 +13,7 @@ import { renderToStaticMarkup } from "react-dom/server";
 import {
   OLLAMA_PLACEHOLDER_SECRET,
   supportedCredentialProviders,
-} from "@workbench/hub-client/credential-test";
+} from "@workbench/connections/credential-test";
 
 import { App } from "../src/app";
 import { NavigationProvider } from "../src/navigation";
@@ -62,7 +62,7 @@ const signedIn: SessionState = {
 describe("CREDENTIAL_PROVIDERS", () => {
   // ProviderPicker (onboarding-page.tsx) renders one card per entry here —
   // this is the data the cards are built from, so covering it covers what
-  // actually shows up: one card per `@workbench/hub-client` provider the
+  // actually shows up: one card per `@workbench/connections` provider the
   // hub can actually test a credential against, each with a distinct
   // honest one-liner and a real key-console link. Compared against
   // `supportedCredentialProviders()` directly, not a second hand-copied