Skip to content

OT Patch and Vulnerability Management with CISA Advisories

Run OT vulnerability management that works: read CISA ICS advisories, pull CSAF data, check KEV, rank by exposure and consequence, then patch or compensate.

Published on 8 min read

In IT, vulnerability management usually means scan, rank by CVSS, patch within a deadline. In operational technology that loop breaks at almost every step. Active scanning can upset fragile controllers. Patches need vendor qualification and an outage window that may be months away. Some devices will never get a fix because the product line is end of life. And a CVSS score says nothing about whether the vulnerable PLC runs a cooling water pump or a safety interlock.

None of that means OT vulnerabilities are unmanageable. It means the process has to be built around asset knowledge, vendor advisories and consequence, with compensating controls treated as a normal outcome rather than a failure. This guide walks through a practical workflow using the public data CISA publishes.

Start from the inventory, not the scanner

You cannot match advisories to assets you do not know about. Before anything else you need, for each device: vendor, product, model, firmware or software version, location (zone and Purdue level), and the process function it supports. Passive monitoring and controlled native queries are the safest ways to build that list; OT asset inventory with passive monitoring covers the techniques.

Firmware version is the field that matters most and is most often missing. A PLC listed as "controller, line 3" cannot be matched to an advisory that affects firmware versions before a certain release. Make version a required field and refresh it after every maintenance window.

Reading CISA ICS advisories

CISA publishes ICS advisories for vulnerabilities in industrial products, usually in coordination with the vendor or the researcher who found the issue. They are listed on the CISA ICS advisories page and follow a stable naming scheme:

  • ICSA-YY-DDD-NN: an ICS advisory, where YY is the year, DDD the day of the year it was released, and NN a sequence number for that day. For example ICSA-26-323-01 was released on day 323 of 2026.
  • ICSMA- prefixes the same scheme for medical device advisories.
  • A trailing letter (ICSA-10-316-01a) marks an update to an earlier advisory.

Each advisory follows roughly the same structure: a summary with CVSS scores, affected products and versions, vulnerability details with CVE and CWE identifiers, background (critical infrastructure sectors, countries deployed), and mitigations. The mitigation section is the part to read carefully. It often contains the vendor's fixed version alongside workarounds such as disabling a service, restricting a port or changing a configuration setting, and it usually repeats CISA's standing recommendations: minimize network exposure, keep control networks behind firewalls and isolated from business networks, use secure remote access methods.

Getting advisories as data: CSAF

Reading web pages does not scale beyond a handful of vendors. CISA also publishes its advisories in the Common Security Advisory Framework (CSAF) 2.0, an OASIS standard JSON format, in the public cisagov/CSAF GitHub repository. OT advisories live under csaf_files/OT/white/<year>/, and the folder includes an index.txt listing every file and a changes.csv with the latest modification times.

Pull the newest advisories:

BASE=https://raw.githubusercontent.com/cisagov/CSAF/develop/csaf_files/OT/white
curl -s "$BASE/changes.csv" | head -20

Summarize one advisory with jq:

curl -s "$BASE/2026/icsa-26-323-01.json" | jq '{
  id:      .document.tracking.id,
  title:   .document.title,
  vendor:  .product_tree.branches[0].name,
  cves:    [.vulnerabilities[].cve],
  fixes:   [.vulnerabilities[].remediations[]? | select(.category=="vendor_fix") | .details]
}'

CSAF documents describe affected products in a product_tree with vendor, product name and version branches, and each vulnerability lists its CVE, scores and remediations (categories such as vendor_fix, workaround and mitigation). That structure lets you script the first match: take the vendors in your inventory, check every new advisory's product_tree for them, and open a ticket when one appears. Treat automated matching as a filter for human review, because product naming in advisories does not always match how devices identify themselves on the network.

Check the Known Exploited Vulnerabilities catalog

CISA's Known Exploited Vulnerabilities (KEV) catalog lists CVEs with evidence of active exploitation. It covers both IT and OT products, and it is published as JSON:

curl -s https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json \
  | jq -r '.vulnerabilities[]
      | select(.vendorProject | test("Schneider|Siemens|Rockwell|Honeywell|Emerson|ABB"; "i"))
      | [.cveID, .vendorProject, .product, .dateAdded] | @tsv'

Replace the vendor list with the vendors from your own inventory (the names above are only examples). Do not forget the Windows, VPN, firewall and remote access products that sit in Levels 3 and 3.5: in practice, much exploitation that reaches OT starts with those IT components rather than with a controller.

Prioritize by exposure and consequence

CVSS measures properties of the vulnerability. OT prioritization needs two more inputs:

  1. Exposure. Can an attacker reach the vulnerable interface? A flaw in a web server on a PLC that is reachable only from one engineering workstation in a well segmented zone is very different from the same flaw on a device reachable from the enterprise network or the internet. Your segmentation design is a direct input here.
  2. Consequence. What happens to the process if this device is compromised or crashes? Map each asset to the worst credible outcome: safety, environmental, equipment damage, production loss, or none.

A simple triage that holds up in practice:

PriorityConditionTypical response
Act nowIn KEV, or exploit public, and the asset is reachable from outside its zoneMitigate within days (block, disable service, isolate), patch at next feasible window
PlannedReachable only inside its zone, high consequence assetPatch in the next scheduled outage, add monitoring meanwhile
TrackLow exposure and low consequence, or no fix existsDocument, apply compensating controls, review at each refresh

CISA's Stakeholder Specific Vulnerability Categorization (SSVC) offers a more formal decision tree using exploitation status, technical impact and mission impact. It maps well to OT because "mission impact" is where process consequence enters.

Patching safely

When you do patch, the IEC 62443 series has specific guidance: the technical report ISA-TR62443-2-3 (Part 2-3) covers patch management in the IACS environment, including the information exchange between product suppliers and asset owners. In practice:

  • Use only vendor qualified patches. For DCS and SCADA platforms, the control system vendor typically tests operating system and application updates before approving them. Applying an unqualified Windows update to an HMI can break the control software.
  • Test first. A test bench or staging system that mirrors production lets you install the update, restart, and confirm communication with controllers before touching the plant.
  • Verify integrity. Download from the vendor portal, compare the published hash, and move the file into OT through a controlled path (see secure remote access for staging patterns).
Get-FileHash .\firmware_v4.2.1.bin -Algorithm SHA256
  • Back up before changing. Take a fresh copy of controller programs and configurations, and image the workstations, so you can roll back.
  • Coordinate with operations. Firmware updates often require the controller to go to program or stop mode, which stops the process it controls. Schedule them with the outage plan and redundancy switchovers.
  • Re-baseline after. Update the inventory with the new firmware version and capture new golden images.

Compensating controls when you cannot patch

Many OT vulnerabilities stay open for a long time. That is acceptable only when the residual risk is understood and reduced. Common compensating controls:

  • Restrict network reachability of the vulnerable service to the few hosts that need it, with conduit firewall rules.
  • Disable unused services on the device (embedded web servers, FTP, SNMP write, unused protocols) where the vendor supports it.
  • Use controller protections: physical key switch in run mode, write protection, protection passwords for project and online access, where the product provides them.
  • Add monitoring for exploitation attempts and for engineering actions on the asset, using an OT aware network sensor. The techniques to watch are covered in MITRE ATT&CK for ICS.
  • Plan replacement for end of life products that will never be fixed, and record the decision and the owner.

Record every compensating control against the vulnerability, with a review date. Auditors and insurers increasingly ask for this, and it helps when the next advisory for the same product arrives.

Avoid active scanning surprises

Authenticated or unauthenticated vulnerability scans with IT defaults have crashed PLCs and network modules. If you scan at all in OT, do it in a maintenance window, against a test system first, with OT safe scan profiles, low rates and only the protocols you understand. For most Level 1 and Level 2 devices, matching the passively observed firmware version against advisories gives you the same answer without the risk.

Putting it into a routine

A sustainable cadence:

  • Daily or weekly: pull new CSAF advisories and KEV additions, auto match against inventory vendors, triage matches.
  • Monthly: review the open vulnerability list, compensating controls, and upcoming outage windows.
  • Per outage: apply qualified patches, re-baseline, close tickets.
  • Yearly: review end of life products and replacement plans.

This routine is also what NIST SP 800-82 Rev. 3 expects under its flaw remediation and patch management guidance; see the NIST SP 800-82 overview. And when a vulnerability turns out to have been exploited in your environment, hand over to your OT incident response plan.

References

Related articles

Build an OT incident response plan around safety and process continuity: roles, decision rights, playbooks, evidence from PLCs and HMIs, and verified recovery.
Design vendor remote access to OT you can approve, watch and revoke: brokered sessions, DMZ jump hosts, MFA, time limits and firewall rules to enforce it.
Design OT network segmentation that holds up: zones, an industrial DMZ, adjacent-level firewall rules, ICS protocol allow lists, data diodes and safe rollout.