‹ Back to apps
promptguard
LLM security for AI coding agents -- protect applications from prompt injection, PII leakage, and data exfiltration. Works with Cursor, Claude Code, Codex, Copilot, Windsurf, and any MCP client.
What’s inside
1 MCP server · 4 skills- MCP serverpromptguardPromptGuard's hosted MCP server. Sign in with your PromptGuard account over OAuth; no API key.
- Skillsecure-llm-integrationAdd PromptGuard security to any LLM-powered application. Use when building new AI features, integrating LLM SDKs, or securing existing unprotected LLM calls. Supports Python, Node.js, and TypeScript projects using OpenAI, Anthropic, Google, Cohere, or AWS Bedrock.
- Skillpromptguard-scanScan the current project for LLM SDK calls PromptGuard does not protect, hardcoded API keys, and PromptGuard misconfiguration, then report them by severity and offer fixes. Use when the user asks to run promptguard-scan, or to audit a project for unprotected LLM calls or leaked keys. Reads only; changes nothing until the user agrees.
- Skillpromptguard-secureAdd PromptGuard to the current project now: install promptguard-sdk, add init() at the entry point, set PROMPTGUARD_API_KEY, and verify. Use when the user asks to run promptguard-secure, to add PromptGuard to this project, or to fix what promptguard-scan found. To compare integration methods, use secure-llm-integration instead.
- Skillpromptguard-apiGenerate correct PromptGuard API calls for scanning, redacting, and guarding LLM traffic. Use when writing code that calls PromptGuard endpoints, building integrations, or when the user asks about PromptGuard scan, redact, guard, or proxy APIs.