test: increase WebDriverIO timeout for LLM-judged E2E tests

Increase global mocha timeout from 60s to 180s in wdio.conf.js to accommodate longer-running LLM judge tests that evaluate agent responses and code generation. Add explicit per-test overrides for Phase B scenarios B4 and B5 to ensure adequate time for agent startup, execution, and LLM verification.

- wdio.conf.js: global timeout 60_000 → 180_000ms
- phase-b.test.ts: explicit 180_000ms timeout for B4 and B5 scenarios

v2/tests/e2e/specs/phase-b.test.ts

@@ -220,6 +220,7 @@ describe('Scenario B4 — LLM-Judged Agent Response', () => {
   const SKIP_MSG = 'Skipping — LLM judge not available (no CLI or API key)';
 
   it('should send prompt and get meaningful response', async function () {
+    this.timeout(180_000); // agent needs time to start + run + respond
     if (!isJudgeAvailable()) {
       console.log(SKIP_MSG);
       this.skip();
@@ -300,6 +301,7 @@ describe('Scenario B5 — LLM-Judged Code Generation', () => {
   const SKIP_MSG = 'Skipping — LLM judge not available (no CLI or API key)';
 
   it('should generate valid code when asked', async function () {
+    this.timeout(180_000); // agent needs time to start + run + respond
     if (!isJudgeAvailable()) {
       console.log(SKIP_MSG);
       this.skip();

v2/tests/e2e/wdio.conf.js

@@ -103,7 +103,7 @@ export const config = {
   framework: 'mocha',
   mochaOpts: {
     ui: 'bdd',
-    timeout: 60_000,
+    timeout: 180_000,
   },
 
   // ── Reporter ──