Skip to content
(787)-476-7950 Hormigueros, PR 00660

Mon – Fri 7:00 am – 5:00 pm OPEN Saturday-Sunday CLOSED

Glowing shield with a padlock over a purple circuit board, representing cybersecurity protection

Guide

Cybersecurity Baseline: Which Controls to Put in Place First, and in What Order

A cybersecurity baseline is the short list of controls aimed at the most common attacks, put in place before anything more advanced is bought. For a small or mid-size organization a sensible order is: name an owner, inventory devices and accounts, require multi-factor authentication, patch automatically, keep isolated backups proven by a real restore, protect every endpoint, filter and segment the network, harden email, keep administrator rights off everyday accounts, turn on logging, train staff and write an incident plan. The list maps to the six functions of NIST CSF 2.0 and builds on the 56 Safeguards of CIS Controls v8.1 Implementation Group 1.

Updated November 2026

The three public references behind a baseline

Three public references overlap on what a baseline should cover:

  • NIST Cybersecurity Framework (CSF) 2.0, published on February 26, 2024, sorts security outcomes into six functions: Govern, Identify, Protect, Detect, Respond and Recover. NIST SP 1300 is its quick-start guide for small businesses.
  • CIS Critical Security Controls v8.1, released in June 2024, has 18 Controls. Its Implementation Group 1 (IG1), 56 Safeguards that CIS calls essential cyber hygiene, is meant for organizations with limited IT and security expertise.
  • CISA Cybersecurity Performance Goals (CPG) 2.0, released in December 2025, set baseline goals such as MFA with the strongest method available (phishing-resistant first) and backups tested at least once a year.

None of the three is mandatory on its own, but contracts, insurers and regulators may refer to them.

Steps 1 to 6: know what you have, lock it and patch it

Each step names the CSF 2.0 function it serves and the CIS v8.1 Safeguards involved.

  • 1. Owner and obligations (Govern). One person, with a backup, decides on security and knows which laws, contracts and insurance terms apply.
  • 2. Inventory (Identify). Devices, software and accounts (1.1, 2.1, 5.1). What is not on the list is not patched, protected or restored.
  • 3. MFA (Protect). IG1 requires it for externally exposed applications, remote network access and administrator accounts (6.3 to 6.5).
  • 4. Automatic patching (Identify and Protect). Operating systems and applications, monthly or more often (7.3, 7.4). Rank the backlog with CISA's Known Exploited Vulnerabilities (KEV) catalog, the list of flaws already exploited in real attacks, and patch internet-facing systems first.
  • 5. Isolated, tested backups (Protect and Recover). Automated backups at least weekly plus an isolated copy (11.2, 11.4). CIS leaves the restore test (11.5) to IG2, but CPG 2.0 asks for one at least yearly, so build it in from the start.
  • 6. Endpoint protection (Protect and Detect). Anti-malware on every asset (10.1) and a host firewall on servers and user computers (4.4, 4.5).

Steps 7 to 11: contain, watch and rehearse

  • 7. Filtering and segmentation (Protect). An edge firewall blocks unsolicited inbound traffic, and separate networks for staff, guests, cameras and servers keep one infected device from reaching the rest. CIS covers segmentation in Safeguard 12.2 (IG2) and CPG 2.0 in goal 3.I; the UniFi network planning guide shows a VLAN and guest layout.
  • 8. Email and web browsing (Protect). DNS filtering on user devices (9.2), plus SPF, DKIM and DMARC records for every domain that sends mail. CPG 2.0 goal 3.L asks for DMARC set to “reject”; CIS lists DMARC (9.5) in IG2. Publish “none” first with a reporting address (rua) and move to “reject” only when the reports show that every legitimate sender passes.
  • 9. Least privilege (Protect). Admin work only from a separate privileged account (5.4), and accounts idle for 45 days disabled (5.3).
  • 10. Logging (Protect and Detect). A log management process, logging switched on and enough storage (8.1 to 8.3). IG2 adds retention of at least 90 days (8.10) and weekly review (8.11).
  • 11. Training and incident plan (Protect, Respond and Recover). Training at hire and at least yearly, including social engineering (14.1, 14.2); an incident handler with a backup, a contact list and a way for staff to report (17.1 to 17.3). Keep a printed and an offline copy of the plan, as CISA's #StopRansomware Guide advises.

MFA, EDR and MDR: what they are and when they pay off

  • MFA (multi-factor authentication) adds a second proof to the password. CISA's fact sheet Implementing Phishing-Resistant MFA (October 2022) ranks FIDO/WebAuthn authenticators (hardware keys or keys built into a laptop or phone) and PKI smart cards as phishing-resistant; app or token codes and number-matching push come next, push without number matching after that, and SMS or voice last, as a temporary option only. Worth it: always.
  • EDR (endpoint detection and response) is an agent on each computer and server that records activity, flags suspicious behavior rather than only known signatures, and can isolate the device remotely. CIS names EDR as an example for Safeguard 13.7 (host-based intrusion prevention), an IG3 item. Worth it once patching is under control and someone reads its alerts.
  • MDR (managed detection and response) is a service, not a product: an outside team watches EDR and other alerts, usually around the clock, investigates them and can contain a device or account. Worth it when nobody inside can answer an alert at night or on a weekend.

SIEM and vulnerability scanning: when to add them

  • SIEM (security information and event management) gathers logs from firewalls, servers, cloud email and identity systems in one place and raises alerts on suspicious patterns. Worth it when there are several log sources and someone reviews the alerts; a SIEM nobody watches only stores data. CIS puts centralized alerting (13.1) in IG2.
  • Vulnerability scanning checks systems automatically for missing patches and weak settings. CIS IG2 asks for internal scans quarterly or more often (7.5) and scans of internet-facing assets monthly or more often (7.6). Worth it once patching is automated: the scan proves the patches landed and finds devices missing from the inventory.

A sensible sequence: the eleven steps, then EDR with someone answering its alerts (in-house or MDR), then scanning, then a SIEM.

Field notes: security controls at storm- and outage-prone sites

Where storms, heat and power cuts are routine, security controls fail in ways that are easy to miss:

  • Clocks. Devices that restart without network time stamp their logs wrongly, which breaks an incident timeline. Sync every device to the same time sources (CIS Safeguard 8.4, IG2, asks for at least two).
  • Improvised internet. When the main link or the firewall is down, someone may plug a consumer router or a phone hotspot into the office network, with no filtering or logs. Agree on an approved backup connection in advance.
  • MFA lockout. Phones run flat and text messages stop arriving in a long outage. Give every administrator a second method, such as a spare hardware key or recovery codes kept in a safe.
  • Storm-themed phishing. Fake relief, insurance, delivery and payment messages follow a disaster. Warn staff before the season and remind them how to report one.
  • Paper copies. Email, VPN and file shares may be down when the plan is needed: keep the plan, the contacts and the restore steps on paper too.

The first hour of an incident: a checklist

Based on the response checklist in CISA's #StopRansomware Guide, which asks for the first three steps in that order:

  1. Isolate the affected systems: unplug the network cable or disconnect Wi-Fi. If several systems or subnets look affected, take the network offline at the switch.
  2. Power devices down only if they cannot be disconnected, because that destroys evidence held in memory.
  3. Triage: decide which critical systems are restored first, on a clean network.
  4. Coordinate by phone or another out-of-band channel; the attacker may be reading company email and chat.
  5. Start a written timeline: what was seen, when, by whom and what was done.
  6. Check that the offline backup is untouched, and keep it disconnected.
  7. Keep the ransom note and the logs, and wipe nothing yet.
  8. Call the people in the plan (IT or security provider, cyber insurer, legal counsel) and report to the authorities of your country; in the United States, CISA or the FBI.
  9. Do not pay or contact the attacker without advice. The guide's authors (CISA, the FBI, the NSA and MS-ISAC) advise against paying: it does not ensure decryption, an end to the compromise or that the data stays unpublished.

Mass password resets come later, once affected systems are cleaned and rebuilt. NIST SP 800-61 Rev. 3 (April 2025) covers the full response cycle, organized by CSF 2.0 function.

Frequently Asked Questions

Is antivirus enough, or is EDR needed?

Classic antivirus matches files against known malware. CPG 2.0 goal 4.A also asks for behavior, heuristic or anomaly detection, which current antivirus with behavior-based detection can provide (CIS puts it in IG2 as Safeguard 10.7). EDR goes further, with a history of each device and remote isolation, but it only adds value when someone reviews its alerts, so decide who will do that before buying it.

What is the difference between EDR and MDR?

EDR is the tool installed on computers and servers. MDR is a service in which an outside security team monitors that tool and related alerts, investigates and takes the first containment steps. An organization with its own security staff on call may run EDR alone; one without anyone to answer alerts after hours usually needs MDR. The managed IT services guide explains how to compare outside providers.

How often should backups be tested?

CISA CPG 2.0 asks for backup and recovery tests at least once a year, and CIS Safeguard 11.5 (IG2) asks for recovery tests on a sample of assets quarterly or more often. A useful test restores a whole critical system, not a single file, and times it. File sync services are not backups: encrypted files can sync over the good copies.

Should a small organization follow NIST CSF or CIS Controls?

Both, for different jobs. NIST CSF 2.0 frames what to achieve and suits reports to management, insurers and customers; CIS Controls IG1 is the concrete task list. Organizations that hold or seek U.S. Department of Defense contracts may also face CMMC, which builds on federal contract safeguarding rules and NIST SP 800-171; see the CMMC readiness guide.

About this guide

Author: GlobalTech Corp., a company based in Hormigueros, Puerto Rico, with over 25 years of experience. It serves organizations in Puerto Rico and in Florida, where the work is remote support plus on-site visits for the projects that call for them. It is a SonicWall Certified MSSP for Puerto Rico and runs firewalls, endpoint protection and MFA for its clients. Its compliance work stops at readiness support: it does not assess or certify anyone. Every standard, figure and date in this guide was rechecked at its source in November 2026. Related pages: GlobalTech cybersecurity services and cybersecurity support for Puerto Rico businesses.

More guides: Technology Guides.