OpenAI GPT-5 improvements (#9)

A series of fixes for making GPT-5 reliable and effective with anyclaude.

## UX changes

- Converts opaque OpenAI service errors into 429s so CC natively retries instead of failing. These are very common.
- Rich debugging support via `ANYCLAUDE_DEBUG`
- Supports specifying reasoning effort and service tier 

## Codebase

- Added CI

## Remaining issues

- GPT-5 often fails to use the native tool calls. These failures seems intermittent.

.github/workflows/ci.yml

@@ -0,0 +1,47 @@
+name: CI
+
+on:
+  push:
+    branches: [ main, master ]
+  pull_request:
+    branches: [ main, master ]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    
+    steps:
+    - uses: actions/checkout@v4
+    
+    - name: Setup Bun
+      uses: oven-sh/setup-bun@v2
+      with:
+        bun-version: latest
+    
+    - name: Cache dependencies
+      uses: actions/cache@v4
+      with:
+        path: |
+          ~/.bun/install/cache
+          node_modules
+        key: ${{ runner.os }}-bun-${{ hashFiles('**/bun.lockb') }}
+        restore-keys: |
+          ${{ runner.os }}-bun-
+    
+    - name: Install dependencies
+      run: bun install --frozen-lockfile
+    
+    - name: Run tests
+      run: bun test
+    
+    - name: Type check
+      run: bun run typecheck
+    
+    - name: Build
+      run: bun run build
+    
+    - name: Verify build output
+      run: |
+        test -f dist/main.js
+        test -x dist/main.js
+        head -n 1 dist/main.js | grep -q "#!/usr/bin/env node"