UNPKG

@tosin2013/mcp-shrimp-task-manager

Version:

Enhanced MCP Shrimp Task Manager with comprehensive LLM integration. A task management tool built for AI Agents, emphasizing chain-of-thought, reflection, and style consistency. Features real GPT-4 ↔ MCP tools communication, comprehensive testing pipeline

48 lines (47 loc) 1.65 kB
/** * Data models for workflow state management * * These interfaces define the structure of workflow state and related data * for the Idea Honing Tool, supporting continuity when developers pause and resume work. */ import { CodebaseAnalysisResult } from './analysis-result.js'; /** * Tracks the state of the specification creation workflow * * Workflow state enables developers to pause and resume work on a specification, * preserving context and progress across sessions. */ export interface WorkflowState { /** Associated specification identifier */ specId: string; /** Current phase of the workflow */ currentPhase: 'analysis' | 'drafting' | 'review' | 'planning'; /** IDs of completed sections */ completedSections: string[]; /** Questions that need answers from the user */ pendingQuestions: Question[]; /** Analysis results if the analysis phase is completed (optional) */ analysisResults?: CodebaseAnalysisResult; /** Timestamp of the last activity */ lastActive: Date; } /** * Represents a question to be asked during the interactive process * * Questions guide users through the specification creation process, * ensuring all necessary information is collected. */ export interface Question { /** Question identifier */ id: string; /** Question text */ text: string; /** Context explaining why the question is being asked */ context: string; /** Section that the question relates to */ targetSection: string; /** Whether the question has been answered */ answered: boolean; /** The provided answer (optional) */ answer?: string; }