@macaw-ai/client 0.0.0-stage → 0.9.11

package/payload/secureAI/macaw_adapters/anthropic_ext_tools/secure_anthropic.py

@@ -0,0 +1,840 @@
+#!/usr/bin/env python3
+"""
+SecureAnthropic - MACAW-protected Anthropic Claude wrapper.
+
+Drop-in replacement for Anthropic client with MACAW security.
+Uses MAPL-compliant tool names: tool:<app_name>/generate, tool:<app_name>/complete.
+
+Supports three usage paths:
+1. Direct on service: client.messages.create() - simplest, app-level identity
+2. bind_to_user: service.bind_to_user(user_client) - per-user identity for SaaS
+3. invoke_tool: user.invoke_tool("tool:xxx/generate", ...) - explicit A2A control
+
+Architecture (service mode):
+- macaw_client: Primary client owning LLM operations (tool:{app_name}/generate, etc.)
+- _tools_client: Internal client owning user-registered tools (tool:{app_name}-tools/*, not externally reachable)
+- server_id: Points to macaw_client (the primary interface)
+- Tool isolation: Only macaw_client can invoke tools on _tools_client
+
+Install with: pip install -e /path/to/secureAI
+"""
+
+import os
+import json
+import logging
+import inspect
+from typing import Dict, Any, Optional, List
+
+from anthropic import Anthropic
+from anthropic.types import Message, ContentBlock, TextBlock, Usage
+from macaw_client import MACAWClient
+
+logger = logging.getLogger(__name__)
+
+
+class _StreamContextManager:
+    """
+    Context manager that provides Anthropic-compatible streaming interface.
+
+    Usage:
+        with client.messages.stream(...) as stream:
+            for text in stream.text_stream:
+                print(text, end="")
+    """
+
+    def __init__(self, iterator):
+        self._iterator = iterator
+        self._text_chunks = []
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        # Clean up iterator if needed
+        pass
+
+    @property
+    def text_stream(self):
+        """Yield just the text content from streaming chunks."""
+        for chunk in self._iterator:
+            if isinstance(chunk, dict):
+                # Handle content_block_delta events
+                if chunk.get('type') == 'content_block_delta':
+                    delta = chunk.get('delta', {})
+                    text = delta.get('text', '')
+                    if text:
+                        yield text
+            elif hasattr(chunk, 'delta') and hasattr(chunk.delta, 'text'):
+                # Handle raw event objects
+                if chunk.delta.text:
+                    yield chunk.delta.text
+
+    def __iter__(self):
+        """Iterate over raw chunks."""
+        return iter(self._iterator)
+
+
+class SecureAnthropic:
+    """
+    Drop-in replacement for Anthropic Claude client with MACAW security.
+
+    This version uses a two-client architecture for tool isolation:
+    - macaw_client: Primary, owns LLM operations (tool:{app_name}/generate, etc.)
+    - _tools_client: Internal, owns user-registered tools (not externally reachable)
+
+    server_id points to macaw_client, making it the primary external interface.
+    Tools are isolated: only macaw_client can invoke tools on _tools_client.
+
+    Supports two modes:
+    1. Service mode (default): Creates both clients, handles Claude calls
+    2. User mode (jwt_token provided): Creates user agent with identity for direct calls
+
+    For multi-user scenarios, use service mode + bind_to_user() to create per-user wrappers.
+    """
+
+    def __init__(
+        self,
+        api_key: str = None,
+        app_name: str = None,
+        intent_policy: Optional[Dict[str, Any]] = None,
+        # User mode parameters
+        jwt_token: str = None,
+        user_name: str = None
+    ):
+        """
+        Initialize SecureAnthropic wrapper.
+
+        Args:
+            api_key: Anthropic API key (or from env)
+            app_name: Application name for registration
+            intent_policy: Application-defined MACAW intent policy
+            jwt_token: If provided, creates user agent with this identity (user mode)
+            user_name: Optional user name for user mode
+        """
+        # Get API key
+        self.api_key = api_key or os.environ.get('ANTHROPIC_API_KEY')
+        if not self.api_key:
+            raise ValueError("Anthropic API key required")
+
+        # Real Claude client
+        self.claude_client = Anthropic(api_key=self.api_key)
+
+        # Application identity
+        self.app_name = app_name or "secure-claude-app"
+
+        # Determine mode based on jwt_token
+        self._mode = "user" if jwt_token else "service"
+        self._jwt_token = jwt_token
+        self._user_name = user_name
+
+        # User-registered tools (Claude can call these)
+        self.user_tools = {}
+
+        # Auto-discovered tools cache
+        self._discovered_tools = {}
+        self._tools_registered = False
+
+        # Application-provided intent policy (no defaults!)
+        self.intent_policy = intent_policy or {}
+
+        if self._mode == "service":
+            # SERVICE MODE: Two-client architecture for tool isolation
+            # - macaw_client: Primary, owns LLM operations (externally reachable)
+            # - _tools_client: Internal, owns user tools (not externally reachable)
+
+            # Tools client name (for trace clarity)
+            self._tools_name = f"{self.app_name}-tools"
+
+            # LLM tools on the primary client
+            llm_tools = {
+                f"tool:{self.app_name}/generate": {
+                    "handler": self._handle_generate,
+                    "prompts": ["messages"]  # messages param is a prompt
+                },
+                f"tool:{self.app_name}/complete": {
+                    "handler": self._handle_complete,
+                    "prompts": ["prompt"]  # prompt param is a prompt
+                }
+            }
+
+            # Primary client - owns LLM operations, externally reachable via server_id
+            self.macaw_client = MACAWClient(
+                app_name=self.app_name,
+                app_version="1.0.0",
+                intent_policy=self.intent_policy,
+                tools=llm_tools
+            )
+
+            # Internal tools client - owns user-registered tools, NOT externally reachable
+            self._tools_client = MACAWClient(
+                app_name=self._tools_name,
+                app_version="1.0.0",
+                intent_policy=self.intent_policy,
+                tools={}  # User tools added via register_tool
+            )
+        else:
+            # USER MODE: Create user agent with identity
+            self._tools_name = None
+            self._tools_client = None
+            self.tools = {}
+
+            self.macaw_client = MACAWClient(
+                user_name=user_name,
+                iam_token=jwt_token,
+                agent_type="user",
+                app_name=self.app_name,
+                app_version="1.0.0",
+                intent_policy=self.intent_policy or {
+                    "purpose": f"Claude access for {user_name or 'user'}"
+                }
+            )
+
+        # Register with LocalAgent and get server ID(s)
+        if self._mode == "service":
+            # Register both clients
+            if self.macaw_client.register() and self._tools_client.register():
+                self.server_id = self.macaw_client.agent_id  # Primary interface!
+                self._tools_server_id = self._tools_client.agent_id  # Internal, not exposed
+                logger.info(f"SecureAnthropic registered (two-client mode):")
+                logger.info(f"   Primary (LLM): {self.server_id}")
+                logger.info(f"   Internal (tools): {self._tools_server_id}")
+            else:
+                raise RuntimeError("Failed to register with MACAW LocalAgent")
+        else:
+            # User mode - single client
+            if self.macaw_client.register():
+                self.server_id = self.macaw_client.agent_id
+                logger.info(f"SecureAnthropic registered as: {self.server_id} (mode: {self._mode})")
+            else:
+                raise RuntimeError("Failed to register with MACAW LocalAgent")
+
+        # Mimic Anthropic API structure
+        # MACAW-protected namespaces (route through PEP)
+        self.messages = self._MessagesNamespace(self)
+        self.completions = self._CompletionsNamespace(self)
+
+    # =========================================================================
+    # Pass-through properties for non-MACAW-protected APIs
+    # These delegate directly to the underlying Anthropic client
+    # =========================================================================
+
+    @property
+    def beta(self):
+        """
+        Pass-through to Anthropic beta API (message batches).
+
+        Not MACAW-protected - batch operations use service's API key directly.
+        """
+        return self.claude_client.beta
+
+    def count_tokens(self, *args, **kwargs):
+        """
+        Pass-through to Anthropic count_tokens method.
+
+        Not MACAW-protected - token counting uses service's API key directly.
+        """
+        return self.claude_client.count_tokens(*args, **kwargs)
+
+    def bind_to_user(self, user_client: 'MACAWClient') -> 'BoundSecureAnthropic':
+        """
+        Bind this SecureAnthropic service to a user's MACAW client.
+
+        Creates a lightweight wrapper that routes calls through the user's
+        client to this service, enabling per-user identity and policy enforcement.
+
+        Only valid in service mode.
+
+        Security validations:
+        - Service must be in service mode
+        - User client must be a valid MACAWClient instance
+        - User client must be registered with LocalAgent
+        - User client should be a "user" type agent
+
+        Args:
+            user_client: A registered MACAWClient with user identity
+
+        Returns:
+            BoundSecureAnthropic wrapper for this user
+
+        Raises:
+            ValueError: If service not in service mode or client validation fails
+        """
+        if self._mode != "service":
+            raise ValueError("bind_to_user() only valid for service-mode SecureAnthropic")
+
+        # Validate user_client is a MACAWClient instance
+        if not hasattr(user_client, 'agent_id') or not hasattr(user_client, 'invoke_tool'):
+            raise ValueError("bind_to_user() requires a valid MACAWClient instance")
+
+        # Validate user_client is registered
+        if not getattr(user_client, 'registered', False):
+            raise ValueError("bind_to_user() requires a registered MACAWClient. Call register() first.")
+
+        # Validate agent_type is "user" (warning only, not blocking)
+        agent_type = getattr(user_client, 'agent_type', None)
+        if agent_type and agent_type != "user":
+            logger.warning(f"bind_to_user() called with agent_type='{agent_type}' (expected 'user'). "
+                          f"User identity and policy enforcement may not work as expected.")
+
+        return BoundSecureAnthropic(self, user_client)
+
+    def register_tool(self, name: str, handler: callable) -> 'SecureAnthropic':
+        """
+        Register a tool that Claude can call.
+
+        Tools are registered on _tools_client (internal, not externally reachable).
+        This provides tool isolation - only the LLM can invoke these tools.
+
+        Args:
+            name: Tool name (must match Claude function name)
+            handler: Function to handle tool execution
+
+        Returns:
+            Self for chaining
+        """
+        # Store with MAPL-compliant name using tools client's app_name
+        mapl_name = f"tool:{self._tools_name}/{name}"
+
+        # Wrap handler to accept params dict and unpack as kwargs
+        def wrapped_handler(params):
+            return handler(**params)
+
+        self.user_tools[name] = wrapped_handler  # Store by simple name for lookup
+        self.user_tools[mapl_name] = wrapped_handler  # Also store by MAPL name
+
+        # Register with _tools_client (internal) using MAPL name
+        self._tools_client.register_tool(mapl_name, wrapped_handler)
+
+        logger.info(f"Registered tool: {name} -> {mapl_name}")
+        return self
+
+    def _auto_discover_tools(self, tools_metadata):
+        """
+        Auto-discover tool implementations from caller's context.
+
+        Args:
+            tools_metadata: List of tool definitions from Claude call
+        """
+        if self._tools_registered:
+            return  # Already discovered
+
+        # Get caller's frame (need to go back through the call stack)
+        frame = inspect.currentframe()
+        caller_frame = frame.f_back.f_back.f_back
+        if not caller_frame:
+            logger.warning("Could not access caller frame for auto-discovery")
+            return
+
+        caller_globals = caller_frame.f_globals
+        caller_locals = caller_frame.f_locals
+
+        # Combine locals and globals for search
+        search_scope = {**caller_globals, **caller_locals}
+
+        # Discover implementations for each tool
+        for tool_def in tools_metadata:
+            func_name = tool_def.get("name")
+
+            # Use MAPL-compliant name with tools client: tool:<app_name>-tools/<func_name>
+            mapl_name = f"tool:{self._tools_name}/{func_name}"
+            if func_name and func_name not in self.user_tools:
+                # Look for function in caller's scope
+                if func_name in search_scope:
+                    func = search_scope[func_name]
+                    if callable(func):
+                        logger.info(f"Auto-discovered tool: {func_name} -> {mapl_name}")
+                        self._discovered_tools[func_name] = func
+                        self.user_tools[func_name] = func
+                        self.user_tools[mapl_name] = func
+                else:
+                    logger.warning(f"Tool '{func_name}' not found in caller's scope")
+
+        # Sync all discovered tools with _tools_client (internal)
+        if self._discovered_tools:
+            # Register each discovered tool with _tools_client
+            for func_name, func in self._discovered_tools.items():
+                mapl_name = f"tool:{self._tools_name}/{func_name}"
+                # Create wrapper handler that accepts dict parameter
+                def create_handler(tool_func):
+                    def handler(params):
+                        return tool_func(**params)
+                    return handler
+                self._tools_client.register_tool(mapl_name, create_handler(func))
+            self._tools_registered = True
+            logger.info(f"Auto-registered {len(self._discovered_tools)} tools with _tools_client")
+
+    def _handle_generate(self, params: Dict[str, Any]) -> Dict[str, Any]:
+        """
+        Handle 'generate' tool (messages creation).
+
+        This runs on macaw_client. Tool callbacks route to _tools_client (internal).
+        Returns a serializable dict for MACAW protocol, or an iterator for streaming.
+        """
+        try:
+            # Check if streaming is requested
+            is_streaming = params.get('stream', False)
+
+            if is_streaming:
+                # Streaming mode: return iterator
+                return self._stream_generate(params)
+
+            # Non-streaming mode: existing behavior
+            # Call real Claude API (params are clean - no MACAW internals)
+            response = self.claude_client.messages.create(**params)
+
+            # Check if Claude wants to call tools
+            if hasattr(response, 'content') and response.content:
+                for content_block in response.content:
+                    if content_block.type == 'tool_use':
+                        func_name = content_block.name
+                        tool_args = content_block.input
+                        # Convert to MAPL-compliant name using tools client: tool:<app_name>-tools/<func_name>
+                        mapl_name = f"tool:{self._tools_name}/{func_name}"
+                        logger.info(f"Claude requested tool use: {func_name} -> {mapl_name}")
+
+                        # Check if we have this tool (by simple name or MAPL name)
+                        if func_name not in self.user_tools and mapl_name not in self.user_tools:
+                            result = {
+                                'error': f"Tool not found: {mapl_name}",
+                                'available_tools': list(self.user_tools.keys())
+                            }
+                        else:
+                            # Invoke tool through MACAW (goes through PEP!)
+                            # Route from macaw_client to _tools_client (internal)
+                            # This provides tool isolation: only LLM can call tools
+                            try:
+                                result = self.macaw_client.invoke_tool(
+                                    tool_name=mapl_name,
+                                    parameters=tool_args,
+                                    target_agent=self._tools_server_id  # Internal tools client
+                                )
+                                logger.info(f"Tool {func_name} executed successfully via _tools_client")
+                            except Exception as e:
+                                error_msg = str(e)
+                                # Check if MACAW blocked it
+                                if 'denied' in error_msg.lower() or 'policy' in error_msg.lower():
+                                    logger.warning(f"MACAW blocked tool {mapl_name}: {error_msg}")
+                                    result = {
+                                        'error': 'Access denied by security policy',
+                                        'tool': mapl_name,
+                                        'reason': 'Policy violation'
+                                    }
+                                else:
+                                    result = {'error': str(e)}
+
+                        # Build messages for continuation
+                        messages = list(params.get('messages', []))
+
+                        # Add assistant message with tool use (use original func_name for Claude)
+                        messages.append({
+                            "role": "assistant",
+                            "content": [{"type": "tool_use", "id": content_block.id,
+                                        "name": func_name, "input": tool_args}]
+                        })
+
+                        # Add tool result
+                        messages.append({
+                            "role": "user",
+                            "content": [{
+                                "type": "tool_result",
+                                "tool_use_id": content_block.id,
+                                "content": json.dumps(result)
+                            }]
+                        })
+
+                        # Continue conversation with tool results
+                        final_params = params.copy()
+                        final_params['messages'] = messages
+
+                        # Remove tools to prevent infinite recursion
+                        if 'tools' in final_params:
+                            del final_params['tools']
+
+                        final_response = self.claude_client.messages.create(**final_params)
+
+                        # Convert to serializable dict for MACAW protocol
+                        return self._response_to_dict(final_response)
+
+            # No tool calls, convert response to dict
+            return self._response_to_dict(response)
+
+        except Exception as e:
+            logger.error(f"Error in generate handler: {e}")
+            return {'error': str(e)}
+
+    def _stream_generate(self, params: Dict[str, Any]):
+        """
+        Handle streaming message creation.
+
+        Returns an iterator that yields chunks from the Anthropic streaming API.
+        Each chunk is converted to a serializable dict.
+        """
+        try:
+            # Call Anthropic with streaming
+            with self.claude_client.messages.stream(**{k: v for k, v in params.items() if k != 'stream'}) as stream:
+                for event in stream:
+                    # Convert event to serializable dict
+                    if hasattr(event, 'type'):
+                        chunk_dict = {
+                            'type': event.type,
+                        }
+                        # Handle different event types
+                        if event.type == 'content_block_delta':
+                            if hasattr(event, 'delta') and hasattr(event.delta, 'text'):
+                                chunk_dict['delta'] = {'text': event.delta.text}
+                            if hasattr(event, 'index'):
+                                chunk_dict['index'] = event.index
+                        elif event.type == 'message_start':
+                            if hasattr(event, 'message'):
+                                chunk_dict['message'] = {
+                                    'id': event.message.id,
+                                    'model': event.message.model,
+                                    'role': event.message.role,
+                                }
+                        elif event.type == 'message_delta':
+                            if hasattr(event, 'delta'):
+                                chunk_dict['delta'] = {}
+                                if hasattr(event.delta, 'stop_reason'):
+                                    chunk_dict['delta']['stop_reason'] = event.delta.stop_reason
+                            if hasattr(event, 'usage'):
+                                chunk_dict['usage'] = {
+                                    'output_tokens': event.usage.output_tokens
+                                }
+                        elif event.type == 'message_stop':
+                            pass  # Just mark end of stream
+
+                        yield chunk_dict
+
+        except Exception as e:
+            logger.error(f"Error in streaming generate: {e}")
+            # Yield error as final chunk
+            yield {'error': str(e)}
+
+    def _handle_complete(self, params: Dict[str, Any]) -> Dict[str, Any]:
+        """
+        Handle 'complete' tool (legacy completions).
+
+        Returns a serializable dict for MACAW protocol.
+        """
+        try:
+            response = self.claude_client.completions.create(**params)
+            return {
+                'id': getattr(response, 'id', None),
+                'type': 'completion',
+                'completion': response.completion,
+                'model': getattr(response, 'model', None),
+                'stop_reason': getattr(response, 'stop_reason', None)
+            }
+        except Exception as e:
+            logger.error(f"Error in complete handler: {e}")
+            return {'error': str(e)}
+
+    def _response_to_dict(self, response) -> Dict[str, Any]:
+        """Convert Anthropic Message response to serializable dict."""
+        result = {
+            'id': response.id,
+            'type': response.type,
+            'role': response.role,
+            'model': response.model,
+            'content': [],
+            'stop_reason': response.stop_reason,
+            'stop_sequence': response.stop_sequence,
+            'usage': {
+                'input_tokens': response.usage.input_tokens,
+                'output_tokens': response.usage.output_tokens
+            } if response.usage else None
+        }
+
+        # Convert content blocks
+        for block in response.content:
+            if hasattr(block, 'text'):
+                result['content'].append({
+                    'type': 'text',
+                    'text': block.text
+                })
+            elif hasattr(block, 'type') and block.type == 'tool_use':
+                result['content'].append({
+                    'type': 'tool_use',
+                    'id': block.id,
+                    'name': block.name,
+                    'input': block.input
+                })
+
+        return result
+
+    def _dict_to_message(self, data: Dict[str, Any]) -> Message:
+        """Reconstruct Anthropic Message from dict."""
+        # Build content blocks
+        content = []
+        for block_data in data.get('content', []):
+            if block_data.get('type') == 'text':
+                content.append(TextBlock(type='text', text=block_data['text']))
+            # Add other block types as needed
+
+        # Build usage
+        usage = None
+        if data.get('usage'):
+            usage = Usage(
+                input_tokens=data['usage']['input_tokens'],
+                output_tokens=data['usage']['output_tokens']
+            )
+
+        return Message(
+            id=data['id'],
+            type=data.get('type', 'message'),
+            role=data.get('role', 'assistant'),
+            model=data['model'],
+            content=content,
+            stop_reason=data.get('stop_reason'),
+            stop_sequence=data.get('stop_sequence'),
+            usage=usage
+        )
+
+    # Namespace classes to mimic Anthropic API structure
+    class _MessagesNamespace:
+        def __init__(self, parent):
+            self.parent = parent
+
+        def create(self, **kwargs):
+            """
+            Create message - routes through MACAW.
+
+            This mimics Anthropic's API but goes through PEP!
+            Adapter creates authenticated prompts based on its own prompts declaration.
+            Supports streaming when stream=True is passed.
+            """
+            # Auto-discover tools if provided
+            tools = kwargs.get('tools')
+            if tools:
+                self.parent._auto_discover_tools(tools)
+
+            tool_name = f"tool:{self.parent.app_name}/generate"
+            is_streaming = kwargs.get('stream', False)
+
+            # Route through MACAW's generate resource
+            # MAPL-compliant: tool:<service>/generate
+            # Authenticated prompts are auto-created by invoke_tool() based on registry
+            result = self.parent.macaw_client.invoke_tool(
+                tool_name=tool_name,
+                parameters=kwargs,
+                target_agent=self.parent.server_id,  # Route to ourselves!
+                stream=is_streaming  # Explicit streaming flag
+            )
+
+            if is_streaming:
+                # Streaming mode: result is a StreamingResultIterator, wrap in Anthropic types
+                return self._wrap_streaming_response(result)
+
+            # Handle errors
+            if isinstance(result, dict) and 'error' in result:
+                raise Exception(f"MACAW error: {result['error']}")
+
+            # Reconstruct Message from dict
+            return self.parent._dict_to_message(result)
+
+        def _wrap_streaming_response(self, iterator):
+            """Wrap streaming chunks for Anthropic-compatible iteration."""
+            for chunk in iterator:
+                if isinstance(chunk, dict):
+                    if 'error' in chunk:
+                        raise Exception(f"MACAW streaming error: {chunk['error']}")
+                    yield chunk
+                else:
+                    yield chunk
+
+        def stream(self, **kwargs):
+            """
+            Stream message creation - context manager style.
+
+            Usage:
+                with client.messages.stream(...) as stream:
+                    for text in stream.text_stream:
+                        print(text, end="")
+            """
+            # Use create with stream=True, wrap in context manager
+            kwargs['stream'] = True
+            return _StreamContextManager(self.create(**kwargs))
+
+    class _CompletionsNamespace:
+        def __init__(self, parent):
+            self.parent = parent
+
+        def create(self, **kwargs):
+            """Create completion - routes through MACAW."""
+            tool_name = f"tool:{self.parent.app_name}/complete"
+
+            # MAPL-compliant: tool:<service>/complete
+            # Authenticated prompts are auto-created by invoke_tool() based on registry
+            result = self.parent.macaw_client.invoke_tool(
+                tool_name=tool_name,
+                parameters=kwargs,
+                target_agent=self.parent.server_id  # Route to ourselves!
+            )
+
+            # Handle errors
+            if isinstance(result, dict) and 'error' in result:
+                raise Exception(f"MACAW error: {result['error']}")
+
+            # Return raw completion dict (legacy API)
+            return result
+
+
+class BoundSecureAnthropic:
+    """
+    Per-user wrapper for SecureAnthropic service.
+
+    Created via SecureAnthropic.bind_to_user(user_client).
+    Routes all calls through the user's MACAWClient to the service,
+    enabling per-user identity and policy enforcement.
+
+    Authenticated prompts are auto-created by invoke_tool based on
+    the service's prompts declaration in the registry.
+
+    Call unbind() to invalidate this binding when done.
+    """
+
+    def __init__(self, service: SecureAnthropic, user_client: 'MACAWClient'):
+        """
+        Initialize bound wrapper.
+
+        Args:
+            service: The shared SecureAnthropic service (must be in service mode)
+            user_client: User's registered MACAWClient with identity
+        """
+        self._service = service
+        self._user_client = user_client
+        self._bound = True
+
+        # Create Anthropic-compatible API namespaces
+        self.messages = self._MessagesNamespace(self)
+        self.completions = self._CompletionsNamespace(self)
+
+    @property
+    def service(self) -> SecureAnthropic:
+        """Get the bound service (raises if unbound)."""
+        if not self._bound:
+            raise RuntimeError("BoundSecureAnthropic has been unbound. Create a new binding with bind_to_user().")
+        return self._service
+
+    @property
+    def user_client(self) -> 'MACAWClient':
+        """Get the bound user client (raises if unbound)."""
+        if not self._bound:
+            raise RuntimeError("BoundSecureAnthropic has been unbound. Create a new binding with bind_to_user().")
+        return self._user_client
+
+    def unbind(self):
+        """
+        Unbind this wrapper, invalidating all future calls.
+
+        After unbinding:
+        - All API calls will raise RuntimeError
+        - References to service and user_client are cleared
+        - A new binding must be created via bind_to_user()
+
+        This is useful for cleanup and ensuring bindings are not reused
+        after a user session ends.
+        """
+        if not self._bound:
+            logger.warning("BoundSecureAnthropic.unbind() called on already unbound instance")
+            return
+
+        logger.info(f"Unbinding SecureAnthropic from user {self._user_client.agent_id}")
+        self._bound = False
+        self._service = None
+        self._user_client = None
+
+    @property
+    def is_bound(self) -> bool:
+        """Check if this wrapper is still bound."""
+        return self._bound
+
+    # =========================================================================
+    # Pass-through properties for non-MACAW-protected APIs
+    # These delegate to the service's underlying Anthropic client
+    # =========================================================================
+
+    @property
+    def beta(self):
+        """Pass-through to Anthropic beta API (message batches) via service."""
+        return self.service.claude_client.beta
+
+    def count_tokens(self, *args, **kwargs):
+        """Pass-through to Anthropic count_tokens method via service."""
+        return self.service.claude_client.count_tokens(*args, **kwargs)
+
+    class _MessagesNamespace:
+        def __init__(self, bound: 'BoundSecureAnthropic'):
+            self.bound = bound
+
+        def create(self, **kwargs):
+            """
+            Create message via user's client -> service.
+
+            Authenticated prompts are auto-created by invoke_tool.
+            Supports streaming with stream=True parameter.
+            """
+            tool_name = f"tool:{self.bound.service.app_name}/generate"
+            is_streaming = kwargs.get('stream', False)
+
+            # Route through user's client (auto-creates auth prompts!)
+            result = self.bound.user_client.invoke_tool(
+                tool_name=tool_name,
+                parameters=kwargs,
+                target_agent=self.bound.service.server_id,
+                stream=is_streaming  # Explicit streaming flag
+            )
+
+            if is_streaming:
+                # Streaming mode: result is a StreamingResultIterator, wrap for iteration
+                return self._wrap_streaming_response(result)
+
+            # Handle errors
+            if isinstance(result, dict) and 'error' in result:
+                raise Exception(f"MACAW error: {result['error']}")
+
+            # Reconstruct Message from dict
+            return self.bound.service._dict_to_message(result)
+
+        def _wrap_streaming_response(self, iterator):
+            """Wrap streaming chunks for Anthropic-compatible iteration."""
+            for chunk in iterator:
+                if isinstance(chunk, dict):
+                    if 'error' in chunk:
+                        raise Exception(f"MACAW streaming error: {chunk['error']}")
+                    yield chunk
+                else:
+                    yield chunk
+
+        def stream(self, **kwargs):
+            """
+            Stream message creation - context manager style.
+
+            Usage:
+                with user_client.messages.stream(...) as stream:
+                    for text in stream.text_stream:
+                        print(text, end="")
+            """
+            # Use create with stream=True, wrap in context manager
+            kwargs['stream'] = True
+            return _StreamContextManager(self.create(**kwargs))
+
+    class _CompletionsNamespace:
+        def __init__(self, bound: 'BoundSecureAnthropic'):
+            self.bound = bound
+
+        def create(self, **kwargs):
+            """Create completion via user's client -> service."""
+            tool_name = f"tool:{self.bound.service.app_name}/complete"
+
+            result = self.bound.user_client.invoke_tool(
+                tool_name=tool_name,
+                parameters=kwargs,
+                target_agent=self.bound.service.server_id
+            )
+
+            if isinstance(result, dict) and 'error' in result:
+                raise Exception(f"MACAW error: {result['error']}")
+
+            # Return raw completion dict (legacy API)
+            return result