XGuard — Agent Execution Gateway

io.github.moelayyan90Published by GitHub account moelayyan90v5.2.0Updated Nov 7, 2026

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

The publisher's own server: Execution Market

VerifiedStreamable HTTPWeb executableSecurity & MonitoringAI & ML

Overview

AI-generated overview

A remote gateway that runs scoped agent executions with server-side credentials, policy, budgets and verifiable receipts.

What you can do
XGuard is a remote streamable-HTTP endpoint at api.xguardgate.com described as a scoped agent execution gateway. Its stated purpose is to run agent actions under server-side credentials, policy and budgets, and to return verifiable receipts. The registry description lists no tools, so the concrete operations are not documented here. The README mostly describes a separate public facts service, XGuard Live, which publishes source-verified facts over plain HTTP.
When to use it
Consider it when an assistant needs to perform actions through a managed gateway that keeps credentials server-side and enforces policy and spending limits, rather than holding keys locally. It is not the same as the public facts service described in the README, which needs no account and is not an MCP tool set.
Requirements
A remote MCP client that supports streamable HTTP, plus network access to api.xguardgate.com. The manifest declares no authentication, environment variables or headers, and no packages. The README's development and operations sections describe Node 24, Cloudflare Workers and an owner key for admin, but those are for running the service, not for connecting to it.
Before you install
The manifest declares no authentication, so verify how access is actually controlled before sending anything sensitive. The README states that the former execution gateway is retained only for state and recovery compatibility and is not the public Live product, and that historical result recovery still requires its original signed quote. Budgets and paid retrieval are mentioned, so confirm what, if anything, can incur charges. No tool list is published, so capabilities cannot be checked in…

Installation

In SourceWeft

  1. Open XGuard — Agent Execution Gateway in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.moelayyan90%2Fxguard-control-plane/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "xguard-agent-execution-gateway": {
      "type": "http",
      "url": "https://api.xguardgate.com/mcp"
    }
  }
}

README

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

sh
curl https://xguardgate.com/fact/npm/react/latest-versioncurl -H 'Accept: text/markdown' https://xguardgate.com/fact/npm/react/latest-versioncurl -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.

sh
npm ci --prefix apps/relaynpm run checknpm testnpm run buildnpm run --prefix apps/relay live:test-runtimenpm 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.

Source: README.md at commit 91c23ae

Tools

6

xguard.fact.get

Read one free published software fact by canonical ID. Returns the primary source, verification status and timestamp. Use xguard.web.fetch for a separate paid fetch of a public URL.

ParametersfactIdrequired

xguard.capabilities

Discover XGuard capabilities

Return the enabled and disabled XGuard tools, their security boundary, and canonical discovery URLs. Free.

Takes no parameters.

xguard.preflight

Preflight a guarded paid request

Free, read-only safety gate for xguard.web.fetch. Validates the public HTTPS target, SSRF policy, DNS reachability and selected payment path without contacting the target or attempting payment. On allow, call xguard.web.fetch directly; it creates the signed quote and returns the mandatory x402 v2 PaymentRequired result.

Takes no parameters.

xguard.pricing.quote

Get a signed execution quote

Get an optional signed quote for xguard.web.fetch. Call xguard.web.fetch directly to receive the price in an HTTP 402 challenge.

Takes no parameters.

xguard.web.fetch

Fetch a public HTTPS resource

Fetch one public HTTPS resource after x402 v2 payment and on-chain settlement verification. Retry the identical MCP call with Payment-Signature and X-XGuard-Quote. A completed retry returns the prior receipt.

Takes no parameters.

xguard.extract

Extract main text, titles, metadata and links from 1–3 public HTTPS pages into one deduplicated, attributed result. No JavaScript rendering, private pages or web search. Up to 6,000 text characters per page; partial source coverage is explicit. Price: 0.012 USDC on Base. Uses x402 v2 HTTP payment headers: retry identical arguments and idempotency key with Payment-Signature and X-XGuard-Quote; the authorization is for the canonical HTTP API resource. No source is fetched before payment.

Parametersurlsrequiredidempotency_keyrequiredmax_amount_atomicrequired

Version history

2
  1. v5.2.0LatestSep 20, 2026
  2. v5.1.1Sep 16, 2026

Score

90% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listedRequired15 pts
  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

    Declare a license in the repository or the registry entry.

Other Execution integrations

Execution Market

market.executionOfficial publisher

Remote MCP endpoint where AI agents publish bounties for real-world tasks and settle them with gasless USDC payments via x402.

Streamable HTTPWeb executableAuth
FinanceCredential requiredOct 24, 2026

ModelPilot Execution Decision Engine

Published by GitHub account andycywu

Remote MCP endpoint that gives an assistant policy-aware execution decisions, verification, cost and trajectory evidence for agent runs.

Streamable HTTPWeb executableAuth
AI & MLCredential requiredOct 28, 2026

AIT Advisory — Enterprise Execution Diagnostic

Published by GitHub account Aramco2025

AIT Advisory's remote MCP server that lets an assistant present a paid executive diagnostic, check company fit, and share purchase and booking routes.

Streamable HTTPWeb executable
Productivity & Workflow5 toolsOct 24, 2026

DLL Execution

Published by GitHub account dllio

A free, read-only remote MCP server that lets an assistant simulate an unsigned Base Mainnet transaction before signing and verify an already-broadcast one.

Streamable HTTPWeb executable
Finance2 toolsOct 24, 2026

Covalt Execution Gateway

Published by GitHub account covaltpay

Remote MCP gateway for CovaltPay, a payment orchestration platform where agents submit proposals for multi-party, milestone-gated payments that humans must…

Streamable HTTPWeb executableAuth
FinanceCredential requiredOct 27, 2026