| title: RelayShield Agentic Attack Surface (AWS Marketplace) | |
| emoji: 🛡️ | |
| colorFrom: blue | |
| colorTo: red | |
| sdk: gradio | |
| sdk_version: 6.20.0 | |
| app_file: app.py | |
| pinned: false | |
| tags: | |
| - mcp-server | |
| - security | |
| - agent | |
| # RelayShield Agentic Attack Surface — AWS Marketplace Edition | |
| This Space is the AWS Marketplace–registered MCP endpoint for RelayShield's "Consumption Security API Bundles" — Bundle D, Agentic Attack Surface. It exposes 5 tools: | |
| - **MCP Server Risk** — typosquat/reputation/registration-age risk check for MCP server URLs. | |
| - **Prompt-Injection Breach Check** — credential breach exposure sourced specifically from prompt-injection attacks against AI agents. | |
| - **Tech Stack CVE Check** — CISA KEV / high-EPSS CVEs targeting a declared AI agent framework or tech stack. | |
| - **Bulk Identity Risk** — hierarchical org + AI-agent-identity risk scoring for a domain and its agent/service-account identities. | |
| - **LLM Credential Exposure (LLMjacking)** — exposed OpenAI/Anthropic/Google/Groq/xAI/Replicate API keys. | |
| ## API key required per call | |
| Each tool call takes your RelayShield API key as an argument. Your key is issued automatically by email when your AWS Marketplace subscription activates — no separate signup step is required. | |
| ## Using this as an MCP tool | |
| Connect your MCP client directly to this Space's MCP endpoint (`/gradio_api/mcp/sse`). | |
| ## Learn more | |
| Full usage instructions, authentication details, and sample requests are included with your AWS Marketplace subscription. | |