prompts-mcp-server

The prompts-mcp-server maintains conversational context and continuity across development sessions, helping AI assistants manage project state, workflows, and tasks.

• Initialize Project (init_prompts / pmcp setup): Scans a project and generates a prompt system (context.md, workflow logs, todos, dev rules, module directories) without overwriting existing files.

• Bootstrap Session (bootstrap): Loads the full context chain (context.md, daily logs, recent activity, summaries, todos, module records) — the first step an AI agent should call when starting a session.

• Clarify Requirements (check_requirements): Runs a 5-point clarity check on a task description and generates follow-up questions instead of guessing when requirements are ambiguous.

• Generate Execution Plan (make_plan): After requirements are confirmed, generates a concrete, actionable plan awaiting user approval.

• Log Dialogs (log_dialog): Records a conversation session into the context chain (daily log, recent activity, summaries, session state) including decisions, code changes, and leftover todos.

• Log & Read Module Changes (log_module / read_module): Records modifications per code module and retrieves historical records — read module history before making changes.

• Manage Todos (update_todos): Add, complete, or remove items in the project's todos.md.

• Skill System (pmcp skill): Manage and apply specialized role-based skills (e.g., analyst, architect, backend) for different tasks.

• Code Version Control: Facilitates commits, pushes, and publishing workflows including privacy audits (pmcp audit, pmcp publish).

• Tool & Scene Recommendation (pmcp tools / pmcp recommend): Discover and recommend appropriate development tools and scenarios for specific contexts.

1. Click on "Install Server".

2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.

3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@prompts-mcp-server Initialize prompts for my project"

That's it! The server will respond to your query, and you can continue using it as needed.

Here is a step-by-step guide with screenshots.