Fail-closed write gate — SQLGuard — Execution Certificate Firewall MCP Resource

Fail-closed write gate is one of 7 resources on the SQLGuard — Execution Certificate Firewall MCP server. Connect the server and your client can attach it as context.

by client

How to attach Fail-closed write gate from your client

• Claude Code SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate run in your project directory
• Claude Desktop SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate ~/Library/Application Support/Claude/claude_desktop_config.json
• Cursor SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate ~/.cursor/mcp.json
• VS Code SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate .vscode/mcp.json
• Zed SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate ~/.config/zed/settings.json
• Windsurf SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate ~/.codeium/windsurf/mcp_config.json
• Cline SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate ~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json
• Gemini CLI SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate ~/.gemini/settings.json
• Grok SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate .mcp.json (in your project root)
• ChatGPT SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate Settings → Connectors → Advanced → Developer mode
• Claude.ai SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate Settings → Connectors → Add custom connector
• LangChain SQLGuard — Execution Certificate Firewall Resource/Fail-closed write gate pip install langchain-mcp-adapters

Other resources on this server

• Agent discover index
• Handshake - what do you need?
• SQLGuard catalog
• Authorize-Before-Mutate Protocol
• Why agents NEED SQLGuard
• llms.txt

SQLGuard — Execution Certificate Firewall also offers 2 prompts — things you invoke, where a resource is context you attach.

Read from the server itself, by connecting to it and calling resources/list on 24 September 2026. A listing does not declare its resources, so asking is the only way to know them, and this is what the server actually offers rather than what its listing claims. Names only are stored; connect the server for each resource's type and contents.