feat: Add configurable max_tokens support to Summarizer

## Configuration Support

Implemented max_tokens configuration support to allow users to control Claude API response length via configuration files.

### Changes

**memdocs/summarize.py**:
- Added `max_tokens` parameter to `Summarizer.__init__()` with default=DEFAULT_MAX_TOKENS (4096)
- Updated `Summarizer.summarize()` to use `self.max_tokens` instead of hardcoded constant
- Enhanced docstrings to document the new parameter
- Stored max_tokens as instance variable for flexible configuration

**memdocs/cli_modules/commands/review_cmd.py**:
- Updated Summarizer instantiation to pass `doc_config.ai.max_tokens`
- Ensures CLI respects configuration file settings

### Configuration Flow

```
.memdocs.yml (max_tokens: 8192)
    ↓
AIConfig schema (default: 4096)
    ↓
Summarizer class (uses configured value)
    ↓
Claude API request
```

### Backward Compatibility

✅ Default parameter value maintains existing behavior (4096 tokens)
✅ All existing code that calls `Summarizer()` without parameters works unchanged
✅ All tests pass (304 passed, 9 skipped)
✅ No breaking changes to API

### Configuration Hierarchy

1. **DEFAULT_MAX_TOKENS = 4096** - Constant used as parameter default
2. **AIConfig.max_tokens = 4096** - Schema default when not specified in config
3. **.memdocs.yml: max_tokens: 8192** - User configuration (overrides schema default)

### Benefits

- Users can now control response length via .memdocs.yml
- Flexibility for different use cases (short summaries vs detailed docs)
- Maintains safe defaults while allowing customization
- Aligns code behavior with documented configuration options

### Example Usage

```python
# Use default (4096 tokens)
summarizer = Summarizer()

# Use custom value
summarizer = Summarizer(max_tokens=8192)

# Via configuration file (.memdocs.yml sets max_tokens: 8192)
# CLI automatically respects config value
```

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <[email protected]>

Author: GeneAI

memdocs/cli_modules/commands/review_cmd.py

@@ -169,7 +169,7 @@ def review(
 
         # Summarize with AI
         with out.spinner("Generating documentation with Claude Sonnet 4.5"):
-            summarizer = Summarizer(model=doc_config.ai.model)
+            summarizer = Summarizer(model=doc_config.ai.model, max_tokens=doc_config.ai.max_tokens)
             doc_index, markdown_summary = summarizer.summarize(context, scope)
 
         out.success("Documentation generated")

memdocs/summarize.py

@@ -25,7 +25,15 @@
 
 
 class Summarizer:
-    """AI-powered documentation summarizer."""
+    """AI-powered documentation summarizer.
+
+    Attributes:
+        api_key: Anthropic API key
+        model: Claude model name
+        max_tokens: Maximum tokens for API responses
+        client: Anthropic API client
+        rate_limiter: Rate limiter for API calls
+    """
 
     PROMPT_TEMPLATE = """You are a technical documentation AI generating machine-readable docs.
 
@@ -94,12 +102,13 @@ class Summarizer:
 
 Generate the YAML now:"""
 
-    def __init__(self, api_key: str | None = None, model: str = "claude-sonnet-4-5-20250929"):
+    def __init__(self, api_key: str | None = None, model: str = "claude-sonnet-4-5-20250929", max_tokens: int = DEFAULT_MAX_TOKENS):
         """Initialize summarizer.
 
         Args:
             api_key: Anthropic API key (defaults to ANTHROPIC_API_KEY env var)
             model: Claude model to use
+            max_tokens: Maximum tokens for Claude API response (default: 4096)
         """
         self.api_key = api_key or os.environ.get("ANTHROPIC_API_KEY")
 
@@ -115,6 +124,9 @@ def __init__(self, api_key: str | None = None, model: str = "claude-sonnet-4-5-2
         except Exception as e:
             raise ValueError(str(e)) from e
 
+        # Store max_tokens
+        self.max_tokens = max_tokens
+
         self.client = anthropic.Anthropic(api_key=self.api_key)
 
         # Initialize rate limiter (50 calls per minute)
@@ -144,7 +156,7 @@ def summarize(self, context: ExtractedContext, scope: ScopeInfo) -> tuple[Docume
         # Call Claude
         response = self.client.messages.create(
             model=self.model,
-            max_tokens=DEFAULT_MAX_TOKENS,
+            max_tokens=self.max_tokens,
             messages=[
                 {
                     "role": "user",