All partners
MelayaKoreshield logo

Melaya x Koreshield. Melaya and Koreshield: check untrusted input, retrieved context and proposed actions before your agents trust them.

Technology Partner

Prompt injection checks for AI agents: Melaya and Koreshield partner to scan customer input and retrieved context, with a mandatory action check before Melaya writes.

// Why it matters

A stronger product on both sides

What this means for Koreshield

Koreshield users get a governed agent runtime around their scans. A Melaya workflow can screen a ticket, check retrieved notes and send a proposed refund for review, so Koreshield decisions land at the exact points where support agents act, not only in a separate console.

What this means for Melaya

Melaya integrates Koreshield to scan untrusted input, retrieved context and proposed tool actions. Koreshield action checks run as a mandatory guard immediately before Melaya writes; input and context scans are available to assigned agents and pipelines. Support and operations agents gain a specialist control without changing models.

// Two-sided value

A partnership, not a listing

What Melaya brings

  • Visual agents and multi-agent pipelines that put a Koreshield scan step before a model reads text or retrieves context.
  • A mandatory Koreshield action check in code before a gated write executes, which the model cannot skip, and fail-closed human approval so a flagged refund or account change waits for a person.
  • A least-privilege Koreshield connector on a scan-only key: input and context scans available to assigned agents and pipelines, no admin tools, plus execution history per run.

What Koreshield brings

  • Three scan endpoints for customer input, retrieved RAG context and proposed tool actions.
  • A decision with a request ID for input scans or a scan ID for context and action scans, plus severity, confidence and indicators, and detect and enforce modes.
  • An evidence console, validation runs and published research on indirect prompt injection in RAG systems.
// Joint use cases

What we ship together

01

Scan each inbound support ticket before a triage agent reads it, and route flagged tickets to a person.

02

Check retrieved help-center articles and CRM notes for injected instructions before an agent drafts a reply.

03

Propose a refund or account change and let Melaya run its mandatory Koreshield action check before the write, then hold it for human approval with the decision attached.

// Next steps

Build with Melaya and Koreshield

Talk to us about bringing this partnership into your stack, or read the announcement for the full story.

// Cookies
Melaya uses a small set of first-party cookies that are strictly necessary to authenticate you, maintain your session, and protect the platform from abuse. We do not use advertising cookies, cross-site trackers, or third-party analytics by default. The full cookie list is in our Privacy Policy.