test(e2e): drive the TUI keyless smoke through the cross-platform PTY harness

The smoke's inline Python pty driver only ran on POSIX (no termios on
Windows). Rebuild every scenario — banner sweep, scripted conversation with
model switch, /skill:, Code Mode overlay, resume failure, and the dsh CLI
suite (default boot, personal overlay, invalid overlay, --resume flag,
source-path prompt) — as marker-gated action lists on pty-harness.ts, which
drives ConPTY via node-pty on Windows and the Python driver elsewhere. The
harness gains configArgs (bins with built-in default configs), prepare
(workspace seeding), and inspect (post-run log assertions); examples/
declares the session-title provider the shipped cordis.yml now mounts.

packages/ui/tui/tests/harness.ts

@@ -31,7 +31,7 @@ export interface TuiHarnessOptions {
   beforeMount?: (session: Session) => void
   cwd?: string | null
   formatCwd?: TuiRuntime['formatCwd']
-  /** Fake-agent creation options; auto-title resolves its target from `provider`/`model`. */
+  /** Fake-agent creation options (`provider`/`model` seed the model selector's initial target). */
   agentOptions?: AgentOptions
   contextWindow?: number
   contextTokens?: number
@@ -94,8 +94,8 @@ export async function createTuiTestHarness<TerminalType extends Terminal, Exit e
   } else {
     await options.configureContext(ctx)
   }
-  // A configureContext may mount the real LlmService (e.g. the auto-title
-  // suites); only fill the advisory-catalog stub when none was provided.
+  // A configureContext may mount the real LlmService; only fill the
+  // advisory-catalog stub when none was provided.
   if (ctx.get('llm') === undefined) {
     ctx.provide('llm', {
       listProviders() {