Skip to content

Munisec | operated on our own agentic platform

Security findings with evidenceyou can check before you fix them.

No commitment. The scope is agreed in writing before day one.

Your customer asked for evidence. The Assessment gives you the map to answer.

Projects and services delivered

Brands and platforms where we have delivered services.

  • Agoda
  • Flickr
  • Hostinger
  • NBA
  • Netflix
  • Notion Labs
  • PlayStation
  • SHEIN
  • Smart Fit
  • Wolt
  • X.com

Where you start

A finding without proof turns into an argument with your team, and the flaw stays open. Consulting, SOC and Agentic Pentest, delivered on our agentic platform: agents do the repetitive work, and a specialist decides what reaches you.

pick what started the conversation

Your customer asked for evidence. The Assessment gives you the map to answer.

  1. Conversation

    A first 30-minute call, free of charge

  2. Consulting

    Security assessment of your cloud environment, on the provider you already use.

  3. You take away

    Prioritized report to answer customer security questionnaires, SOC 2, ISO 27001 and LGPD (Brazil's data protection law)

Scope defined before any proposal

// WHAT STAYS OPEN

Your SaaS went live. But there is one thing.

The product works, customers use it and nothing broke. Three flaws that pass any functional test and sit in the OWASP Top 10.

  1. It went live

    $ git push origin main

    Deploy complete. app.example.com is live.

    Build passed, login works, first customers are in. Working was the proof the team needed.

  2. But there is one thing

    $ curl https://app.example.com/.env

    SUPABASE_SERVICE_ROLE_KEY=eyJ•••

    The .env file was left out of .gitignore and shipped with the build. Supabase's service_role key bypasses RLS and opens the whole database.

  3. This is where we come in

    $ munisec pentest --scope contract.md

    ✗ /.env public in production ✗ invoices table without RLS in Supabase ✗ session cookie without HttpOnly

    Agentic Pentest looks for what working does not show: exposed files, RLS table by table, session cookie flags. Every finding comes with evidence you can reproduce.

Illustrative case. Fictional domain, key and table, simplified output, no customer data.

Three services, one platform

You start with what you need to prove. We grow with you.

Each service states what is in and out of scope, in writing, before day one.

See how a finding reaches you, with evidence and a fix.

See a sample report
soc · three-level triageIllustrative
  1. agentscorrelates sensor eventsL1 Triage
  2. agentsmaps them to MITRE ATT&CKL1 Triage
  3. agentsclassifies and attaches the evidenceL1 Triage
  4. analystconfirms or dismisses, and prioritizesL2 Analysis
  5. analystinvestigates and recommends, when there is an escalationL3 Specialist
  6. youreceive the alert in Slack, Teams, email or webhook
  7. youapprove any action in your environment
  • Lightweight sensors detect and correlate events, mapped to MITRE ATT&CK
  • Alerts in Slack, Teams, email or webhook, reviewed by an analyst
  • We detect and recommend. No automatic action in your environment

Illustrative flow. Simplified steps.

// WHAT YOU GET

Every finding with evidence and a fix

The Assessment report is not a loose list. Each finding says what was found, why it matters and how to confirm the fix worked.

  • Evidence you can reproduce: the command and output behind each finding
  • Priority reviewed by a specialist, from critical to informational
  • A remediation plan for your team to apply, with the command that confirms each fix
  • A map to CIS, SOC 2, ISO 27001 and LGPD, to answer your customer's questionnaire
  • A 1-hour meeting to walk through the findings
Talk to a specialist

No commitment. The scope is agreed in writing before day one.

Assessment report

Lab account

CriticalPublic uploads bucket

Evidence
"IsPublic": true on the app-uploads bucket
Fix
Turn on Block Public Access and serve images through signed URLs
Confirm the fix
$ aws s3api get-public-access-block --bucket app-uploads
References
CIS AWS Foundations · SOC 2 CC6.1 · LGPD art. 46
  • HighCI user with AdministratorAccess
  • MediumCloudTrail not logging every region
Illustrative example. The report is a remediation plan, not a compliance attestation.
// CERTIFICATIONS

Our team is certified

Individual certifications, earned and maintained by each partner in their own name.

The certifications belong individually to the Munisec partners. They do not represent a partnership with, sponsorship by or endorsement from AWS, Linux Foundation, CNCF, HashiCorp, IBM, OffSec or Fortinet. AWS, Kubernetes, Prometheus, Argo, Terraform, OSCP and Fortinet are trademarks of their respective owners, used only to identify the technologies and certifications.

AWS

All certifications, including Professional and Specialty

Kubernetes

All CNCF certifications for Kubernetes

Terraform

HashiCorp

Argo

CNCF

Prometheus

CNCF

OSCP

OffSec, hands-on penetration testing certification

Fortinet

Fortinet, firewall and network security

  • AWS
  • Kubernetes
  • Terraform
  • Linux and containers
  • eBPF
  • MITRE ATT&CK
  • CIS hardening
  • LGPD
  • Multi-agent architecture
  • AI agent security
// BEFORE THE CALL

Questions that come up before the contract

The answers come from each service's scope and the contract clauses.

Do you need write access to my environment?

No. For the Assessment, you create read-only access in your cloud provider and can revoke it whenever you want. Munisec applies no fix: you get the report with the plan and decide what goes into your environment. In the SOC, the sensors detect and recommend, with no automatic action.

Does my data go through an external LLM?

It can, and you know before you sign. The contract names the external LLM providers and the sub-processors, describes what is sent to each one and sets Munisec as the processor of your data under LGPD. A security incident involving that data is reported within three business days.

Can the test take production down?

Agentic Pentest only starts with written authorization, rules of engagement and a window agreed with you. Destructive exploitation, or testing production outside that window, is out of scope. The Assessment tests nothing: it only reads the account configuration through the read-only access.

Does the report count as SOC 2 or ISO 27001 certification?

No. The report is a prioritized remediation plan, with the map to CIS, SOC 2, ISO 27001 and LGPD that helps you answer your customer's questionnaire and walk into the audit knowing what is missing. An attestation or compliance certification is out of scope for every service.

// CONTACT

Start with what you need to prove

Tell us what prompted the conversation: a customer questionnaire, SOC 2, ISO 27001, LGPD (Brazil's data protection law) or an incident. A specialist reads your request and replies by email.

Talk to Munisec

Location

Brasília, DF, Brazil

  • Free initial conversation
  • No commitment

Talk to a specialist

Do not send passwords or vulnerability details. To report a security flaw, write to security@munisec.com.br.

No commitment. The scope is agreed in writing before day one.

Munisec uses these details to reply to your request and prepare a proposal, and keeps them for 12 months after your last submission. The form also records where the contact came from and, to stop bots, uses Cloudflare Turnstile and a code derived from your IP. The data goes through providers outside Brazil, such as Cloudflare and Resend. Details and your rights in the Privacy Policy.