Sutra

Instructions

Architects a custom context system based on a high-level goal (The Architect).
Returns a blueprint of Sutra components (Molecules, Cells, Organs, Thinking Models).

Use this when the user wants to build a persistent agent or complex workflow
rather than solving a single immediate task.

Args:
    goal: The user's objective (e.g., "Build a writing assistant that learns my style").
    constraints: Optional limits (e.g., "Must be lightweight").

Implementation Reference

• src/context_engineering_mcp/server.py:75-140 (handler)

  The core handler function for the design_context_architecture tool. Decorated with @mcp.tool() for MCP registration. Validates inputs and generates a context architecture blueprint based on goal and constraints using heuristic matching.

  @mcp.tool()
  def design_context_architecture(goal: str, constraints: str | None = None) -> dict:
      """
      Architects a custom context system based on a high-level goal (The Architect).
      Returns a blueprint of Sutra components (Molecules, Cells, Organs, Thinking Models).
  
      Use this when the user wants to build a persistent agent or complex workflow
      rather than solving a single immediate task.
  
      Args:
          goal: The user's objective (e.g., "Build a writing assistant that learns my style").
          constraints: Optional limits (e.g., "Must be lightweight").
      """
      try:
          model = DesignArchitectureInput(goal=goal, constraints=constraints)
      except ValidationError as e:
          return {"error": str(e)}
  
      g = model.goal.lower()
      c = (model.constraints or "").lower()
  
      # Blueprint Defaults
      blueprint: dict[str, Any] = {
          "name": "Custom System",
          "rationale": "General purpose context structure.",
          "components": {
              "molecule": "Standard CoT",
              "cell": "cell.protocol.key_value",
              "organ": None,
              "cognitive": "reasoning.understand_question",
          },
      }
  
      if "lightweight" in c:
          blueprint["name"] += " (Light)"
  
      # Heuristic Architecture Logic
      if "debate" in g or "perspective" in g:
          blueprint["name"] = "Debate System"
          blueprint["components"]["organ"] = "organ.debate_council"
          blueprint["rationale"] = "Uses a multi-perspective organ to balance viewpoints."
  
      elif "research" in g or "report" in g or "synthesize" in g:
          blueprint["name"] = "Research Engine"
          blueprint["components"]["organ"] = "organ.research_synthesis"
          blueprint["components"]["cell"] = (
              "cell.protocol.episodic"  # Log research trails
          )
          blueprint["rationale"] = (
              "Combines a synthesis organ with episodic memory to track findings."
          )
  
      elif "learn" in g or "remember" in g or "style" in g:
          blueprint["name"] = "Adaptive Assistant"
          blueprint["components"]["cell"] = "cell.protocol.windowed"
          blueprint["rationale"] = (
              "Uses windowed memory to maintain recent context and style."
          )
  
      elif "code" in g or "bug" in g or "review" in g:
          blueprint["name"] = "Code Auditor"
          blueprint["components"]["cognitive"] = "reasoning.verify_logic"
          blueprint["rationale"] = "Focuses on logic verification for code correctness."
  
      return blueprint

• src/context_engineering_mcp/server.py:63-70 (schema)

  Pydantic input schema (BaseModel) used for validating the parameters of the design_context_architecture tool.

  class DesignArchitectureInput(BaseModel):
      goal: str = Field(
          ..., min_length=5, description="The goal of the system to design."
      )
      constraints: str | None = Field(
          None, description="Optional constraints or preferences."
      )

• src/context_engineering_mcp/server.py:75-75 (registration)

  The @mcp.tool() decorator registers the design_context_architecture function as an MCP tool.

  @mcp.tool()