sync(ci): update workflows and mage-x to v1.10.3 (#75)

Author: mrz1836

.github/.env.base

@@ -137,15 +137,6 @@ TEST_TIMEOUT_RACE_COVER=30m            # Timeout for tests with race+coverage (m
 TEST_TIMEOUT_UNIT=20m                  # Timeout for unit tests only
 TEST_TIMEOUT_FUZZ=5m                   # Timeout for fuzz tests
 
-# ================================================================================================
-# 📡 GO-BROADCAST CONFIGURATION
-# ================================================================================================
-
-# Automerge Labels Configuration
-# When using --automerge flag, these labels will be added to created PRs
-# Comma-separated list of labels to apply for automatic merging
-GO_BROADCAST_AUTOMERGE_LABELS=automerge
-
 # ================================================================================================
 # 📊 GO-COVERAGE SYSTEM CONFIGURATION
 # ================================================================================================
@@ -244,8 +235,9 @@ REDIS_CACHE_FORCE_PULL=false           # Force pull Redis images even when cache
 # 🪄 MAGE-X CONFIGURATION
 # ================================================================================================
 
-MAGE_X_VERSION=v1.8.14                                        # https://github.com/mrz1836/mage-x/releases
+MAGE_X_VERSION=v1.10.3                                        # https://github.com/mrz1836/mage-x/releases
 MAGE_X_USE_LOCAL=false                                        # Use local version for development
+MAGE_X_CI_SKIP_STEP_SUMMARY=true                              # Skip duplicate test results in step summary (already in test validation summary)
 MAGE_X_AUTO_DISCOVER_BUILD_TAGS=true                          # Enable auto-discovery of build tags
 MAGE_X_AUTO_DISCOVER_BUILD_TAGS_EXCLUDE=race,custom           # Comma-separated list of tags to exclude
 MAGE_X_FORMAT_EXCLUDE_PATHS=vendor,node_modules,.git,.idea    # Format exclusion paths (comma-separated directories to exclude from formatting)
@@ -284,7 +276,7 @@ MAGE_X_YAMLFMT_VERSION=v0.20.0                                # https://github.c
 # MAGE_X_DOWNLOAD_TIMEOUT=5000
 # MAGE_X_DOWNLOAD_USER_AGENT=MAGE-X-Agent
 # MAGE_X_PARALLEL=3
-# MAGE_X_TEST_EXCLUDE_MODULES=module1,module2               
+# MAGE_X_TEST_EXCLUDE_MODULES=module1,module2
 # MAGE_X_TEST_RACE=false
 # MAGE_X_VERBOSE=true
 
@@ -458,3 +450,53 @@ PR_MANAGEMENT_SIZE_XS_THRESHOLD=10
 PR_MANAGEMENT_SIZE_S_THRESHOLD=50
 PR_MANAGEMENT_SIZE_M_THRESHOLD=200
 PR_MANAGEMENT_SIZE_L_THRESHOLD=500
+
+# ================================================================================================
+# 📡 GO-BROADCAST CONFIGURATION & AI-POWERED TEXT GENERATION
+# ================================================================================================
+
+# Automerge Labels Configuration
+# When using --automerge flag, these labels will be added to created PRs
+# Comma-separated list of labels to apply for automatic merging
+GO_BROADCAST_AUTOMERGE_LABELS=automerge
+
+# AI generates intelligent PR descriptions and commit messages based on diff analysis.
+# Disabled by default. All AI failures fall back to static templates silently.
+# Uses Google Genkit SDK with support for Anthropic, OpenAI, and Google providers.
+
+# Master switch - enables AI infrastructure (disabled by default)
+GO_BROADCAST_AI_ENABLED=false
+
+# Granular controls (default to GO_BROADCAST_AI_ENABLED value)
+GO_BROADCAST_AI_PR_ENABLED=      # Enable AI for PR body generation
+GO_BROADCAST_AI_COMMIT_ENABLED=  # Enable AI for commit message generation
+
+# Provider: anthropic, openai, google
+GO_BROADCAST_AI_PROVIDER=anthropic
+
+# API key (or use provider-specific: ANTHROPIC_API_KEY, OPENAI_API_KEY, GEMINI_API_KEY)
+# DO NOT USE THIS IN PUBLIC REPOSITORIES, USE ENVIRONMENT SECRETS INSTEAD
+# GO_BROADCAST_AI_API_KEY=
+
+# Model override (uses provider defaults if empty)
+# anthropic: claude-sonnet-4-5-20250929 | openai: gpt-5.2 | google: gemini-3-pro-preview
+GO_BROADCAST_AI_MODEL=
+
+# Generation parameters
+GO_BROADCAST_AI_MAX_TOKENS=2000
+GO_BROADCAST_AI_TIMEOUT=30
+GO_BROADCAST_AI_TEMPERATURE=0.3
+
+# Diff truncation (prevents token limit issues)
+GO_BROADCAST_AI_DIFF_MAX_CHARS=4000
+GO_BROADCAST_AI_DIFF_MAX_LINES_PER_FILE=50
+
+# Response caching (reduces API calls for identical diffs across repos)
+GO_BROADCAST_AI_CACHE_ENABLED=true
+GO_BROADCAST_AI_CACHE_TTL=3600
+GO_BROADCAST_AI_CACHE_MAX_SIZE=1000
+
+# Retry settings (handles transient failures)
+GO_BROADCAST_AI_RETRY_MAX_ATTEMPTS=3
+GO_BROADCAST_AI_RETRY_INITIAL_DELAY=1
+GO_BROADCAST_AI_RETRY_MAX_DELAY=10