Exponencial Administração & Tecnologia
ISO 9001 ISO 14001 ISO 27001
PT Get started

Technology line · Applied AI · Audit

AI tools to protect, govern and audit agents.

Exponencial is turning its own internal agent-security architecture into an open line for companies: API, SDK, skills, GitHub, npm and a continuous audit platform.

Request a demo View tools Products

01 Tools

One line, multiple entry points.

Each tool solves a practical pain and also brings the company into the platform: prompt security, agent governance, ISO evidence, risk triage and real-use inventory.

API · SDK · npm

Agent Shield

Prompt injection protection for chats, RAG, agents and pipelines. Automatic signup, organization API keys, quotas, scheduled batch and UUID feedback.

View Agent Shield

Hooks · Rules · GitHub

Agent Safe Guard

Agent governance at the action point: dangerous-command blocking, decision trace and a rule catalog for technical teams.

View governance

Sec-AI · ISO 27001

AI audit and Sec-AI

Triage of vulnerabilities, risk signals and security evidence to feed ISO 27001, ISO 42001 and AI controls.

View platform

Public skill · A.8.8

ISO A.8.8 + npm audit

Skill that turns npm audit into traceable technical vulnerability evidence, with documented pending items and remediation.

View skill

Inventory · Shadow AI

Agents, skills and extensions

Discovery and governance of AI agents, MCPs, skills, VS Code extensions and privileges already running in operations.

View inventory

Governance · ISO 42001

AI system governance

Policies, risks, controls and evidence to treat AI as a managed, auditable system aligned with the business.

View ISO 42001

Operating model

Protection does not sit in a report. It runs before the decision.

The design is simple: detect risk at the action point, record context, return a UUID, allow customer feedback and turn demo usage into a commercial relationship.

02 Commercial flow

Open demo, clear limits and a quota path.

External companies can start testing without a meeting. When quota ends, the API itself points to authentication, quota request and Exponencial's sales flow.

  1. 01

    Protect before action

    Prompt, command, tool call or batch job goes through an objective barrier before touching environments, customers or sensitive data.

  2. 02

    Record without hiding context

    Each verdict generates a UUID, metadata, input and decision reason for debugging, audit and service improvement.

  3. 03

    Learn from real feedback

    The API accepts customer assessment by UUID: good block, false positive, false negative, too strict or too permissive.

  4. 04

    Turn quota into relationship

    When external usage passes the limit, the response points to authentication, quota request and Exponencial approval.

03 Stack and API

SDKs, skills and repositories as entry points.

npm and GitHub make the line discoverable by technical teams. The site organizes the commercial proposal and routes the customer to demo, quota or assisted deployment.

API

/api/guard/*

signup, screen, batch, results, feedback, stats, billing, quota and organization API keys.

npm

@exponencial/agent-shield

Public JavaScript SDK and CLI for fast technical integration.

GitHub

exponencialadm/agent-shield

Commercial README, examples and agent skill as technical SEO.

Dashboard

Organizations and quota

limit approval, API-key usage and lead capture for assisted deployment.
Agent Shield API Agent Safe Guard ISO A.8.8 skill npm GitHub

Product and pipeline

The product edge is connecting AI to real governance.

Agents, audit and security stop being separate pages: they become a product line for selling technology, capturing leads and refining the service with real usage data.

I want to evaluate it Agent Shield