Model Context Protocol moderate risk

Shrimp Task Manager

Shrimp Task Manager is an MCP server for AI-assisted development, providing persistent memory, structured workflows, and smart task decomposition for efficient project management.

Connections & Capabilities

Connects To

GitHub

Capabilities

readwriteexecadmin

Quickstart

Install

npx -y

Config

{
  "mcpServers": {
    "shrimp-task-manager": {
      "command": "node",
      "args": ["/path/to/mcp-shrimp-task-manager/dist/index.js"],
      "env": {
        "DATA_DIR": "/path/to/your/shrimp_data",
        "TEMPLATES_USE": "en",
        "ENABLE_GUI": "false"
      }
    }
  }
}

Exposed MCP Tools (7)

moderate

init project rules

Initializes project standards by creating or updating configuration files.

Modifies project configuration files.

safe

plan task

Creates a task plan based on a given description, breaking it down into subtasks.

Analyzes requirements and generates a plan without executing code.

high

execute task

Executes a specific task, potentially modifying files and running commands.

Can execute arbitrary code and modify the file system.

high

continuous mode

Executes all tasks sequentially, automating the development workflow.

Automates potentially dangerous operations without human intervention.

safe

list tasks

Shows all tasks and their status.

Read-only operation that displays task information.

safe

research

Enters research mode to explore technologies and solutions.

Gathers information without modifying the system.

moderate

reflect task

Reviews and improves a task based on its previous execution.

May modify task plans based on reflection.

Safety Assessment

Shrimp Task Manager offers significant productivity enhancements but requires careful configuration. It is relatively safe when used with well-defined project rules and close monitoring. Risks increase when used with untrusted code or without proper security measures for the data directory and web GUI.

Requires explicit configuration and initialization.
Task execution can be monitored and controlled.
Project rules can be defined to maintain coding standards.
Optional web GUI can be disabled.

Writes to the file system based on agent decisions.
No built-in sandboxing for task execution.
Relies on the security of the underlying Node.js environment.
Potential for information disclosure if data directory is not properly secured.
If web GUI is enabled, it introduces web-based attack surface.