feat(auth): add OAuth2 support for xAI with PKCE and token persistence

- Implemented xAI OAuth2 integration with PKCE (Proof Key for Code Exchange) support.
- Added logic for token exchange, refresh, and persistent storage in JSON format.
- Created `xai` package with helpers for OAuth discovery, API token handling, and URL building.
- Introduced `XAIExecutor` for integrating xAI credentials into runtime HTTP requests.
- Added unit tests to validate OAuth flow, token persistence, and endpoint validation.

internal/registry/model_definitions.go

@@ -21,6 +21,7 @@ type staticModelsJSON struct {
 	CodexPro    []*ModelInfo `json:"codex-pro"`
 	Kimi        []*ModelInfo `json:"kimi"`
 	Antigravity []*ModelInfo `json:"antigravity"`
+	XAI         []*ModelInfo `json:"xai"`
 }
 
 // GetClaudeModels returns the standard Claude model definitions.
@@ -78,6 +79,11 @@ func GetAntigravityModels() []*ModelInfo {
 	return cloneModelInfos(getModels().Antigravity)
 }
 
+// GetXAIModels returns the standard xAI Grok model definitions.
+func GetXAIModels() []*ModelInfo {
+	return cloneModelInfos(getModels().XAI)
+}
+
 // WithCodexBuiltins injects hard-coded Codex-only model definitions that should
 // not depend on remote models.json updates. Built-ins replace any matching IDs
 // already present in the provided slice.
@@ -167,6 +173,7 @@ func cloneModelInfos(models []*ModelInfo) []*ModelInfo {
 //   - codex
 //   - kimi
 //   - antigravity
+//   - xai
 func GetStaticModelDefinitionsByChannel(channel string) []*ModelInfo {
 	key := strings.ToLower(strings.TrimSpace(channel))
 	switch key {
@@ -186,6 +193,8 @@ func GetStaticModelDefinitionsByChannel(channel string) []*ModelInfo {
 		return GetKimiModels()
 	case "antigravity":
 		return GetAntigravityModels()
+	case "xai", "x-ai", "grok":
+		return GetXAIModels()
 	default:
 		return nil
 	}
@@ -208,6 +217,7 @@ func LookupStaticModelInfo(modelID string) *ModelInfo {
 		data.CodexPro,
 		data.Kimi,
 		data.Antigravity,
+		data.XAI,
 	}
 	for _, models := range allModels {
 		for _, m := range models {