"laya" matching MCP servers:

Enables fully offline, token-free screening of text through a local Laya typed-decision model, exposing prompt-injection/jailbreak detection plus structured choice, score, and yes/no judgments in about 55ms per call. It also reports service status with measured metrics and known limitations.

Enables agents to get fast, calibrated typed decisions from a shared Laya decision service: classifying text, scoring rubrics, answering batches of yes/no checks, and running question packs. Exposes guards and tool-call helpers that gate proposed tool calls, screen tool results for injected instructions, check replies and input for leaks or jailbreaks, and validate individual form fields and whole submissions with pass, fix, review, or reject outcomes.

Enables secure, zero-trust access to MCP tools through short-lived, signed capability leases that bind tool execution to specific sessions, intents, and constraints. Prevents prompt injection attacks and privilege escalation with dynamic risk scoring, policy enforcement, and tamper-evident audit logging.