Custom connections are gone: the CLI talks to the LLeMbas server it is logged in to

LLeMbas CLI 2.0 is LLeMbas-server-specific (CLI rule 2). Removed: the
anthropic, gemini, ollama and responses clients, context discovery and
llama-swap unloading (discover.ts), the Connection schema with its dialect,
auth, key_cmd, discover, unload_url and one_model_at_a_time, the model
`fallback` (it only ever moved to another server), effort_map and cache, and
the multi-connection refs. Kept: openai-chat, http, sse, think, effort,
tokens and all of src/lembas.

- connections.yaml holds the login (`type: webui`, as lembas login writes
  it). A custom entry from 1.x is read past with a warning that says how to
  log in, so an old file still starts; with several logins left, the latest
  is used and the others named.
- A model's ref is the id the server serves it under; the older forms
  (`<login>/<id>`, `<login>/<model>`) still resolve and config.yaml is
  rewritten once. Without a login there is no model, and the error (and the
  first-run screen, now the sign-in alone) says `lembas login <address>`.
- A login to another server signs this machine out of the previous one.
- The window is the server's word (no discovery); a capacity group comes
  from the server's `capacity.group` alone; embeddings use the server's
  /v1/embeddings.
- `lembas models` lists the server's models; install.sh's starter files, the
  README and the JSON schemas say the same.

Tests: every fixture is a login now (the fake server lists models, answers
other GETs with 404, and can hand one model to another script, as a server
with several backends); the dialect suites and the cross-connection tests are
gone with their code.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/lembas-login.test.ts

@@ -187,14 +187,14 @@ test("login writes one webui connection named after the instance, and no models"
 
   const loaded = loadConfig()
   const c = loaded.connections.example!
-  expect(c.dialect).toBe("openai-chat")
+  expect(c.webui?.url).toBe(fake.url)
   expect(c.base_url).toBe(`${fake.url}/v1`)
   expect(c.api_key).toBe("lmb_secret")
   // The instance's models with its settings, the default first; the group named after the connection.
   expect(Object.keys(c.models)).toEqual(["qwen", "gemma"])
   expect(c.models.qwen).toEqual({ name: "Qwen", context: 131072, temperature: 0.6, top_p: 0.95, efforts: ["low", "high"], tools: true, group: "example:gpu" })
   expect(c.models.gemma).toEqual({ context: 32768, vision: true, single_session: true, group: "example:gpu" })
-  expect(resolveModel(loaded, undefined).ref).toBe("example/qwen")
+  expect(resolveModel(loaded, undefined).ref).toBe("qwen")
 })
 
 test("a model changed on the instance is here at the next start, without logging in again", async () => {
@@ -205,7 +205,7 @@ test("a model changed on the instance is here at the next start, without logging
   expect(r).toEqual([{ name: "example", ok: true, models: 1 }])
   const loaded = loadConfig()
   expect(Object.keys(loaded.connections.example!.models)).toEqual(["llama"])
-  expect(resolveModel(loaded, undefined).ref).toBe("example/llama")
+  expect(resolveModel(loaded, undefined).ref).toBe("llama")
   // A model the instance does not list yet is still asked for, not refused here.
   expect(resolveModel(loaded, "example/brand-new").id).toBe("brand-new")
 })
@@ -266,7 +266,7 @@ test("an older kind of login is replaced: the old connection, its key and its pi
   const c = cfg()
   // It was the instance's default, so the instance decides it now; the rest follow the new name.
   expect(c.model).toBeUndefined()
-  expect(c.small_model).toBe("example/gemma")
+  expect(c.small_model).toBe("gemma")
   expect(r.changes.join("\n")).toContain("the connection ai from an earlier login is now example")
 })