Think Strategies

// Plain object schema for MCP SDK export default { type: "object", properties: { // Core properties from original schema thought: { type: "string", description: "Your current thinking step" }, nextThoughtNeeded: { type: "boolean", description: "Whether another thought step is needed" }, thoughtNumber: { type: "integer", minimum: 1, description: "Current thought number" }, totalThoughts: { type: "integer", minimum: 1, description: "Estimated total thoughts needed" }, isRevision: { type: "boolean", description: "Whether this revises previous thinking" }, revisesThought: { type: "integer", minimum: 1, description: "Which thought is being reconsidered" }, branchFromThought: { type: "integer", minimum: 1, description: "Branching point thought number" }, branchId: { type: "string", description: "Branch identifier" }, needsMoreThoughts: { type: "boolean", description: "If more thoughts are needed" }, // Strategy selection and metadata strategy: { type: "string", enum: [ "linear", "chain_of_thought", "react", "rewoo", "scratchpad", "self_ask", "self_consistency", "step_back", "tree_of_thoughts" ], description: "The thinking strategy being employed" }, currentStage: { type: "string", description: "Current stage in the thinking process flow" }, // Strategy-specific properties // ReAct properties action: { type: "string", description: "Specific action to take for gathering information (ReAct)" }, observation: { type: "string", description: "Result or observation from executing an action (ReAct)" }, // ReWOO properties planningPhase: { type: "boolean", description: "Whether currently in planning phase (ReWOO)" }, toolCalls: { type: "array", items: { type: "object", properties: { tool: { type: "string" }, input: { type: "string" } }, required: ["tool", "input"] }, description: "Planned tool calls with inputs (ReWOO)" }, // Scratchpad properties stateVariables: { type: "object", description: "Current state of variables being tracked (Scratchpad)" }, // Self-Ask properties subQuestion: { type: "string", description: "Current sub-question being addressed (Self-Ask)" }, subQuestionAnswer: { type: "string", description: "Answer to the current sub-question (Self-Ask)" }, subQuestionNumber: { type: "integer", minimum: 1, description: "Current sub-question number (Self-Ask)" }, // Self-Consistency properties reasoningPathId: { type: "string", description: "Identifier for current reasoning path (Self-Consistency)" }, pathAnswers: { type: "array", items: { type: "object", properties: { pathId: { type: "string" }, answer: { type: "string" } }, required: ["pathId", "answer"] }, description: "Answers from different reasoning paths (Self-Consistency)" }, // Step-Back properties generalPrinciple: { type: "string", description: "General principle or approach identified (Step-Back)" }, // Tree of Thoughts properties approachId: { type: "string", description: "Identifier for current approach (ToT)" }, approaches: { type: "array", items: { type: "object", properties: { id: { type: "string" }, description: { type: "string" }, promise: { type: "number", minimum: 0, maximum: 10 } }, required: ["id", "description"] }, description: "Different approaches being explored (ToT)" }, evaluationScore: { type: "number", minimum: 0, maximum: 10, description: "Evaluation score for current branch (ToT)" }, // Verification and solution properties hypothesis: { type: "string", description: "Current solution hypothesis" }, verificationResult: { type: "boolean", description: "Whether the hypothesis has been verified" }, verificationReasoning: { type: "string", description: "Reasoning behind verification result" }, finalAnswer: { type: "string", description: "Final verified answer to the problem" }, // Semantic routing properties currentState: { type: "string", description: "Current state in the thinking flow" }, stateDescription: { type: "string", description: "Description of the current state" }, sessionToken: { type: "string", description: "Session identifier token" }, availableActions: { type: "object", additionalProperties: { type: "object", properties: { description: { type: "string" }, requiredInputs: { type: "array", items: { type: "string" } }, optionalInputs: { type: "array", items: { type: "string" } }, hints: { type: "object", additionalProperties: { type: "string" } }, nextState: { type: "string" }, isGlobal: { type: "boolean" } }, required: ["description"] }, description: "Available actions with semantic hints" }, canSwitchStrategy: { type: "boolean", description: "Whether strategy switching is allowed from current state" } }, required: ["thought", "nextThoughtNeeded", "thoughtNumber", "totalThoughts", "strategy"] };