XGuard — Agent Execution Gateway

Official registry AI and agents

Scoped agent execution. Server-side credentials, policy, budgets and verifiable receipts.

Install

Add this to your client's MCP configuration (Claude Desktop: claude_desktop_config.json; Cursor: .cursor/mcp.json; VS Code and others use the same "mcpServers" shape).

Remote (streamable-http)
Claude Code
claude mcp add --transport http xguard-agent-execution-gateway 'https://api.xguardgate.com/mcp'
JSON configuration (Claude Desktop, Cursor, VS Code)
{
    "mcpServers": {
        "xguard-agent-execution-gateway": {
            "type": "http",
            "url": "https://api.xguardgate.com/mcp"
        }
    }
}

Documentation

From the project's README on GitHub, refreshed weekly. Links open on GitHub.

Contents
  1. XGuard Live
  2. Development
  3. Operations
  4. Monetization

XGuard Live

The live facts layer for AI.

Current facts. Source verified. Change tracked. Machine readable.

XGuard Live observes public primary sources and publishes compact, canonical facts with verification timestamps, source evidence and immutable history. People and machines retrieve the same data through ordinary HTTP. No account, SDK, plugin or paid API key is needed for public facts.

Live site · Fact index · Methodology · Changes

curl https://xguardgate.com/fact/npm/react/latest-version
curl -H 'Accept: text/markdown' https://xguardgate.com/fact/npm/react/latest-version
curl -H 'Accept: application/json' https://xguardgate.com/fact/npm/react/latest-version

A published observation includes the current or last observed value, source URL, source hash, observed/verified timestamps, freshness limit, previous value and history. STALE/UNKNOWN never imply a current verified value. Verification means a match to the source at retrieval, not independent certification of every publisher claim. Candidates remain unpublished until schema and source checks pass.

Development

Use Node 24.

npm ci --prefix apps/relay
npm run check
npm test
npm run build
npm run --prefix apps/relay live:test-runtime
npm run dev

Local development uses a separate Worker config and local SQLite Durable Objects. Production smoke checks are read-only and excluded from business metrics: npm run smoke. The primary-source check is node scripts/live-source-check.mjs; robots restrictions, rate limits and changed source schemas can legitimately prevent collection.

Operations

The production Worker is xguard-mainnet. Cron runs every ten minutes; persistent shard alarms collect due sources with domain limits and backoff. Cloudflare cache serves bounded, format-specific representations. /healthz exposes real counts and scheduler health. /admin requires an owner key and protects all mutations.

Deployment: use a pull request, pass CI/CodeQL, merge, then the Deploy XGuard Live workflow deploys the tested commit. Manual deployment with existing Cloudflare credentials: npx --prefix apps/relay wrangler deploy --config apps/relay/wrangler.jsonc --tag git-$(git rev-parse --short=12 HEAD). Production lifecycle changes must first pass node scripts/durable-object-deploy-guard.mjs <deployed-git-ref> HEAD and be isolated from runtime changes.

See operations and rollback and migration inventory. The former execution gateway is retained solely for state/recovery compatibility and historical reference; it is not the public Live product. Historical result recovery still requires its original signed quote. Legacy marketing and release workflows are archived and do not run.

Monetization

Launch mode is observe. Cloudflare Pay Per Crawl eligibility and Pay Per Use buyer participation require confirmation in the owner's Cloudflare account. The application does not fabricate revenue, simulate paid production traffic or equate crawler requests with paid use. Actual provider report imports and invoice costs are separate, auditable records. No report means unknown account revenue, not proven zero. Discovery remains free even when eligible high-value detail pages are later configured for paid retrieval.

Frequently asked questions

What is XGuard — Agent Execution Gateway?

Scoped agent execution. Server-side credentials, policy, budgets and verifiable receipts.

How do I install XGuard — Agent Execution Gateway?

It runs as a remote server. In Claude Code: claude mcp add --transport http xguard-agent-execution-gateway 'https://api.xguardgate.com/mcp'. For other clients, paste the JSON configuration from this page into their MCP settings.

Does XGuard — Agent Execution Gateway need an API key?

Its registry entry lists no required key or secret. Some features may still need an account with the service it connects to.

Is XGuard — Agent Execution Gateway open source?

Its code is public on GitHub, but the repository does not declare a standard license.

Who maintains XGuard — Agent Execution Gateway?

It is published by moelayyan90. AI.tn lists it but does not maintain or endorse it.

From the official MCP Registry. Listing is not an endorsement: review a server's code and permissions before giving it access to your data.