webmcp-sh
Verified Safeby WebMCP-org
Overview
An end-to-end demonstration and playground for the Web Model Context Protocol (WebMCP), allowing users to manage and visualize AI agent memory systems entirely client-side using structured data and SQL.
Installation
npm run devEnvironment Variables
- VITE_SENTRY_DSN
- VITE_SENTRY_ENVIRONMENT
Security Notes
The application runs entirely client-side using PGlite (PostgreSQL in browser) and Drizzle ORM, reducing traditional server-side attack vectors. It implements robust security features for AI agent interaction, including an `analyzeQuery` function in `useMCPSQLTool.ts` that explicitly blocks dangerous SQL operations (e.g., DROP, TRUNCATE, ALTER) and common SQL injection patterns. A protected `audit_log` table with database triggers prevents modification or deletion of audit entries, ensuring integrity. Sentry is integrated for error logging, which collects telemetry data, including potentially PII by default if `sendDefaultPii: true` is left as is, but this is a standard configuration choice. The main security considerations are inherent to browser environments and interaction with external browser extensions.
Similar Servers
memory-graph
A graph-based MCP server that provides intelligent memory capabilities for Claude Code, enabling persistent knowledge tracking, relationship mapping, and contextual development assistance for AI agents.
mcp-duckdb-memory-server
Serves as a local knowledge graph memory backend for LLMs, storing entities, relations, and observations using DuckDB for efficient querying and data management.
context-sync
Providing persistent memory and context synchronization for AI development across various tools and projects.
Simple-Memory-Extension-MCP-Server
A persistent key-value memory store for AI agents, designed to extend context windows and enable semantic search over stored memories.