feat(agents): implement Code Commander module (Phases 1-5)

- Phase 1: Infrastructure (state, prompts, registry)
- Phase 2: Execution engine (AI adapters, security classifier, executors)
- Phase 3: Agent integration (graph nodes, routing)
- Phase 4: Streaming interaction (PTY terminal, WebSocket)
- Phase 5: Frontend integration (Vue components)

backend/app/agents/state.py

@@ -4,7 +4,14 @@ from typing import Annotated, Any, Literal, TypedDict
 
 from app.agents.schemas.event import AgentEvent
 from app.agents.schemas.message import AgentMessage
-from app.agents.schemas.task import AgentTask, CollaborationBudget, InterruptRecord, RecoveryRecord, TaskResult, VerificationStatus
+from app.agents.schemas.task import (
+    AgentTask,
+    CollaborationBudget,
+    InterruptRecord,
+    RecoveryRecord,
+    TaskResult,
+    VerificationStatus,
+)
 from langchain_core.messages import AIMessage, BaseMessage, HumanMessage
 from langgraph.graph.message import add_messages
 
@@ -23,6 +30,7 @@ class AgentRole(str, Enum):
     EXECUTOR = "executor"
     LIBRARIAN = "librarian"
     ANALYST = "analyst"
+    CODE_COMMANDER = "code_commander"
 
 
 @dataclass
@@ -141,6 +149,14 @@ class AgentState(TypedDict):
     user_llm_config: dict[str, Any] | None
     provider_capabilities: dict[str, Any] | None
 
+    # Code Commander state
+    code_task_type: Literal["demo", "project", "modification"] | None
+    code_ai_provider: Literal["claude", "gemini", "codex", "opencode"] | None
+    code_sandbox_mode: bool | None
+    code_workspace_path: str | None
+    code_execution_session_id: str | None
+    code_execution_result: dict[str, Any] | None
+
 
 def initial_state(user_id: str, conversation_id: str) -> AgentState:
     return AgentState(