feat(proxy): add OpenAI Responses API compatibility for Open CoDesign

- Normalize Responses API tool format ({name, parameters} at top level)
  to Chat Completions format ({function: {name, parameters}})
- Handle 'developer' role as system prompt (OpenAI's newer equivalent)
- Add null/non-array guards to convertOpenAIPartsToAnthropicContent and
  extractOpenAITextContent to prevent 'content is not iterable' crashes
- Support input_text, output_text, and reasoning item types in Responses
  API input parsing
- Forward reasoning.effort to Claude adaptive thinking config
- Extend JSON schema sanitization blacklist with draft-2020 keywords
  (patternProperties, , unevaluatedProperties, etc.)
- Add fallback for untyped items with role/content in Responses API input
- Add diagnostic logging for /v1/responses input conversion

src/lib/geminiNexus/JsonSchemaUtils.ts

@@ -162,6 +162,11 @@ function cleanJsonSchemaRecursive(value: any) {
     // 4. Physically remove "hard" blacklist items that interfere with generation
     const hardRemoveFields = [
       '$schema',
+      '$id',
+      '$comment',
+      '$anchor',
+      '$dynamicRef',
+      '$dynamicAnchor',
       'additionalProperties',
       'enumCaseInsensitive',
       'enumNormalizeWhitespace',
@@ -170,7 +175,15 @@ function cleanJsonSchemaRecursive(value: any) {
       'const',
       'examples',
       // Advanced logic fields common in MCP tools but unsupported by Gemini
+      'patternProperties',
       'propertyNames',
+      'additionalItems',
+      'prefixItems',
+      'contains',
+      'minContains',
+      'maxContains',
+      'unevaluatedProperties',
+      'unevaluatedItems',
       'anyOf',
       'oneOf',
       'allOf',