@igor-olikh/openspec-mcp-server 1.1.0 → 1.2.0

package/README.md

@@ -2,8 +2,9 @@
 
 [![NPM Version](https://img.shields.io/npm/v/@igor-olikh/openspec-mcp-server?style=for-the-badge&logo=npm)](https://www.npmjs.com/package/@igor-olikh/openspec-mcp-server)
 [![NPM Downloads](https://img.shields.io/npm/dt/@igor-olikh/openspec-mcp-server?style=for-the-badge&logo=npm)](https://www.npmjs.com/package/@igor-olikh/openspec-mcp-server)
+[![Official Site](https://img.shields.io/badge/Official%20Site-openspec.dev-blue?style=for-the-badge)](https://openspec.dev/)
 
-Welcome! This is a simple bridge (plugin) that connects **[OpenSpec](https://github.com/Fission-AI/OpenSpec)** to your favorite AI coding assistant (like **IBM Bob**, **Codex**, or **Claude Desktop**).
+Welcome! This is a simple bridge (plugin) that connects **[OpenSpec](https://openspec.dev/)** to your favorite AI coding assistant (like **IBM Bob**, **Codex**, or **Claude Desktop**).
 
 ## What is this and why do I need it?
 When you want your AI to build a new feature, you usually just type it into the chat. But as projects grow, the AI can forget things, get confused, or write messy code.

package/dist/server.js

@@ -1,6 +1,6 @@
 import { Server } from '@modelcontextprotocol/sdk/server/index.js';
 import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
-import { CallToolRequestSchema, ErrorCode, ListToolsRequestSchema, McpError, } from '@modelcontextprotocol/sdk/types.js';
+import { CallToolRequestSchema, ErrorCode, ListToolsRequestSchema, ListPromptsRequestSchema, GetPromptRequestSchema, McpError, } from '@modelcontextprotocol/sdk/types.js';
 import { getTools, handleToolCall } from './tools.js';
 export class OpenSpecMCPServer {
     server;
@@ -12,6 +12,7 @@ export class OpenSpecMCPServer {
         }, {
             capabilities: {
                 tools: {},
+                prompts: {},
             },
         });
         this.setupHandlers();
@@ -23,6 +24,42 @@ export class OpenSpecMCPServer {
                 tools: getTools(),
             };
         });
+        this.server.setRequestHandler(ListPromptsRequestSchema, async () => {
+            return {
+                prompts: [
+                    {
+                        name: "openspec_kickoff",
+                        description: "Initialize an AI coding session using Fission-AI OpenSpec best practices.",
+                    }
+                ]
+            };
+        });
+        this.server.setRequestHandler(GetPromptRequestSchema, async (request) => {
+            if (request.params.name !== "openspec_kickoff") {
+                throw new McpError(ErrorCode.InvalidRequest, `Prompt not found: ${request.params.name}`);
+            }
+            return {
+                description: "Initialize an AI coding session using Fission-AI OpenSpec best practices.",
+                messages: [
+                    {
+                        role: "user",
+                        content: {
+                            type: "text",
+                            text: `You are an expert AI development assistant operating within a codebase managed by the Fission-AI OpenSpec framework.
+Your workflow MUST be strictly spec-driven. Never write code until the specification is completely validated and active.
+
+Follow this workflow exactly:
+1. When asked to implement a feature, FIRST check if a change exists by using openspec tools (like \`openspec_status\` or \`openspec_list\`).
+2. If there are no open specs for this feature, create one.
+3. If an active change exists, use tools to read and validate the active specification.
+4. If validation fails, DO NOT PROCEED TO CODE. Fix the spec first.
+5. Once the design is validated and error-free, proceed to implement the feature by tackling the tasks sequentially.
+6. Always communicate your plan clearly to the user.`
+                        }
+                    }
+                ]
+            };
+        });
         this.server.setRequestHandler(CallToolRequestSchema, async (request) => {
             try {
                 const result = await handleToolCall(request.params.name, request.params.arguments || {}, this.projectPath);

package/openspec/changes/add-mcp-prompts/tasks.md

@@ -1,5 +1,5 @@
 # Tasks
 
-- [ ] Implement `ListPromptsRequestSchema` in `src/server.ts`.
-- [ ] Implement `GetPromptRequestSchema` returning the massive text instruction string.
+- [x] Implement `ListPromptsRequestSchema` in `src/server.ts`.
+- [x] Implement `GetPromptRequestSchema` returning the massive text instruction string.
 - [ ] Test in Codex UI using the Quick Action.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@igor-olikh/openspec-mcp-server",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "description": "An MCP server that connects OpenSpec to AI assistants like Codex, Claude, and Cursor.",
   "main": "dist/index.js",
   "type": "module",