claude-prompts
by minipuft
Overview
Manages hot-reloadable AI prompt templates with advanced features like chains, quality gates, and structured reasoning, acting as an automation and code-editing agent for various AI assistants.
Installation
npx -y claude-prompts@latestEnvironment Variables
- NPM_TOKEN
- RELEASE_PLEASE_TOKEN
- DOWNSTREAM_PAT
- MCP_WORKSPACE
- MCP_CONFIG_PATH
- MCP_PROMPTS_PATH
- MCP_METHODOLOGIES_PATH
- MCP_GATES_PATH
- MCP_STYLES_PATH
- LOG_LEVEL
- MCP_LLM_MODEL
- MCP_LLM_ENDPOINT
- MCP_LLM_API_KEY
- MCP_LLM_PROVIDER
- RALPH_SPAWNED
Security Notes
The server includes a 'shell_verify' gate type that explicitly allows executing arbitrary shell commands provided by the user or AI. While it uses `SAFE_ENV_ALLOWLIST` to prevent inheriting sensitive environment variables, the direct execution of unsandboxed, user-defined commands poses a critical security risk if input is not from a fully trusted source or if external sandboxing (e.g., Docker) is not enforced. There are no clear indications of built-in robust sandboxing for these shell executions. Hardcoded secrets are explicitly avoided by design, relying on environment variables for sensitive API keys, and `zod` is used for input validation, which are good practices. However, the `shell_verify` feature inherently lowers the overall safety score for untrusted environments.
Similar Servers
Lynkr
A modular LLM orchestration platform and API proxy, providing a unified interface to various model providers, intelligent routing, and agentic capabilities for complex tasks involving workspace interaction, web browsing, and code development.
conductor-tasks
Conductor Tasks acts as an intelligent AI-powered assistant for developers, streamlining the entire development lifecycle from task generation and planning (parsing PRDs, expanding tasks, generating implementation steps) to execution and code modification (generating diffs). It provides visual task management, integrates with various IDEs, and leverages multiple LLM providers for optimal results and cost efficiency.
responsible-vibe-mcp
Guides AI coding agents through structured development workflows, providing battle-tested engineering methodologies and long-term project memory.
codebase-context
An MCP server providing real-time codebase context (patterns, libraries, architecture, momentum) to AI coding agents to improve code quality and adherence to team standards.