> ## Documentation Index
> Fetch the complete documentation index at: https://yieldxyz.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Security overview

> How yield.xyz keeps every transaction non-custodial and verified: who builds it, how Shield checks it, and why only your keys can execute it.

yield.xyz never holds funds, keys or signing authority. Every action returns an unsigned transaction: we build it, Shield verifies it, and your wallet, custodian or user signs it. Security on the platform is organized around that handoff.

<Card title="Your keys, your crypto" icon="key">
  yield.xyz never has access to user private keys. The APIs only return unsigned transactions for you to sign and broadcast.
</Card>

## The life of a transaction

<Steps>
  <Step title="Craft: we build the transaction">
    Your app requests an [action](/documentation/earn/core-concepts/actions), such as enter, exit or manage. Audited integrations assemble the exact deposit, unstake or redemption, with parameters, routing and gas.
  </Step>

  <Step title="Verify: Shield decodes and checks it">
    [Shield](/documentation/platform/shield) decodes the transaction, simulates the state change and checks it against a verified pattern for that yield: correct contract, function, parameters and receiver. It runs inside the API ([Defense Mode](/documentation/platform/shield/defense-mode)) and, optionally, again in your own stack before signing. Anything that fails is blocked.
  </Step>

  <Step title="Sign: only you can execute it">
    Your app signs with the user's wallet or your custody solution and broadcasts to the chain. yield.xyz is never in this step.
  </Step>
</Steps>

## Controls

<CardGroup cols={2}>
  <Card title="SOC 2 Type II" icon="certificate" href="/documentation/security/audits-compliance">
    Controls are examined continuously, not once. Reports are available on the [Trust Center](https://trust.yield.xyz).
  </Card>

  <Card title="Independent audits" icon="magnifying-glass" href="/documentation/security/audits-compliance#audits">
    Every contract and integration is audited before it goes live.
  </Card>

  <Card title="Live monitoring" icon="satellite-dish">
    Every deployment is watched 24/7. [Sentinel Vaults](/documentation/platform/vaults/sentinel-vaults) add Hypernative and Blockaid monitoring with a pre-signed pause.
  </Card>

  <Card title="On-prem Shield" icon="server" href="/documentation/platform/shield/security-tiers">
    For regulated environments, Shield verification can run inside your own perimeter.
  </Card>

  <Card title="Compliance tooling" icon="globe" href="/documentation/advanced-setup/geo-blocking">
    API-level geoblocking maintained by yield.xyz, based on IP and wallet addresses.
  </Card>

  <Card title="Enterprise SLA" icon="clock">
    99.9% read and 99.5% write availability, up to 1,000 requests per second, and a 30 minute Sev 1 response, 24/7.
  </Card>
</CardGroup>

## What Shield covers

<CardGroup cols={2}>
  <Card title="Shield" icon="shield-check" href="/documentation/platform/shield">
    Decodes every unsigned transaction and validates it against a verified pattern for the specific yield. Runs as a TypeScript library or a standalone binary, so it is usable from any language.
  </Card>

  <Card title="Defense Mode" icon="shield-halved" href="/documentation/platform/shield/defense-mode">
    Shield running server-side, inside the API's own transaction pipeline. A project can run it in Monitor mode (log only) or Block mode (reject failed transactions before they are returned).
  </Card>

  <Card title="HyperNative" icon="bolt" href="/documentation/platform/shield/hypernative">
    A partner integration for pre-signing simulation and protocol monitoring.
  </Card>

  <Card title="Security deployment options" icon="layer-group" href="/documentation/platform/shield/security-tiers">
    How to combine server-side Defense Mode with a client-side check for two independent checkpoints.
  </Card>
</CardGroup>

Shield's checks are structural: it does not evaluate whether the underlying protocol itself is safe, and it does not claim to catch every possible attack. See [Defense Mode](/documentation/platform/shield/defense-mode#threat-model) for what it does and does not mitigate.

## Report a vulnerability

Email [security@yield.xyz](mailto:security@yield.xyz) with a description, steps to reproduce and the potential impact. We acknowledge reports within 24 hours. See the [Security notice](/legal/security-notice#reporting-security-issues) for the full disclosure process.

## Next steps

<CardGroup cols={2}>
  <Card title="Audits & compliance" icon="file-shield" href="/documentation/security/audits-compliance">
    SOC 2 Type II, audit reports and due diligence material.
  </Card>

  <Card title="Trust Center" icon="building-shield" href="https://trust.yield.xyz">
    Security posture and trust artifacts at trust.yield.xyz.
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.