mcp.installers.advanced_code_installer¶

Advanced Code-Generating MCP Server Installer.

Version 2: Uses Aug_LLM agents to generate custom installation code. More flexible but requires human oversight for safety.

Classes¶

`AdvancedCodeInstaller`	Advanced MCP installer with LLM code generation.
`CodeGenerationRequest`	Request for LLM-generated installation code.
`GeneratedInstallPlan`	LLM-generated installation plan.
`SubprocessExecutionInput`	Input for subprocess execution tool.

Module Contents¶

class mcp.installers.advanced_code_installer.AdvancedCodeInstaller(config_manager: haive.mcp.installers.config_manager.MCPConfigManager | None = None)¶

Advanced MCP installer with LLM code generation.

cleanup()¶: Clean up all resources.

create_subprocess_execution_tool(plan: GeneratedInstallPlan, request: CodeGenerationRequest) → langchain_core.tools.StructuredTool¶: Create tool for executing subprocess commands with oversight.

create_validation_tool(plan: GeneratedInstallPlan, request: CodeGenerationRequest) → langchain_core.tools.StructuredTool¶: Create tool for validating installation.

async generate_installation_plan(request: CodeGenerationRequest) → GeneratedInstallPlan¶: Generate installation plan using LLM.

get_advanced_status() → dict[str, Any]¶: Get status of advanced installer.

async install_server_advanced(request: CodeGenerationRequest) → tuple[bool, str, list[langchain_core.tools.StructuredTool]]¶: Advanced server installation with code generation.

code_generation_agent = None¶

config_manager¶

running_servers: dict[str, subprocess.Popen]¶

safe_installer¶

class mcp.installers.advanced_code_installer.CodeGenerationRequest(/, **data: Any)¶

Bases: pydantic.BaseModel

Request for LLM-generated installation code.

context_documents: list[str] = None¶

custom_requirements: str = None¶

package_info: dict[str, Any] = None¶

risk_tolerance: str = None¶

server_description: str = None¶

server_name: str = None¶

class mcp.installers.advanced_code_installer.GeneratedInstallPlan(/, **data: Any)¶

Bases: pydantic.BaseModel

LLM-generated installation plan.

confidence_score: float = None¶

environment_setup: dict[str, str] = None¶

fallback_to_safe: bool = None¶

install_commands: list[str] = None¶

risk_assessment: str = None¶

startup_command: str = None¶

validation_steps: list[str] = None¶

class mcp.installers.advanced_code_installer.SubprocessExecutionInput(/, **data: Any)¶

Bases: pydantic.BaseModel

Input for subprocess execution tool.

command: str = None¶

environment_vars: dict[str, str] = None¶

timeout: int = None¶

working_directory: str | None = None¶