Join our Newsletter — 33% off our NHI Course
Home› Guides› CSAF Security Advisories Guide: Consuming and Publishing Machine-Readable…
Guide Vulnerability Management

CSAF Security Advisories Guide: Consuming and Publishing Machine-Readable Vulnerability Advisories

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 27 October 2026 13 min read
On this page

Vendors still publish most security advisories as web pages and PDFs, so every customer re-types the same product names, versions and fixes into their own tools, slowly and with errors. The Common Security Advisory Framework (CSAF) is the OASIS standard that replaces that with signed, machine-readable JSON advisories that can be matched automatically against an asset inventory or SBOM. This guide explains how CSAF 2.0 documents are structured, which profile to use for which purpose, how distribution and discovery work, how to consume CSAF feeds as a vulnerability management team and how to publish them as a product vendor, including where the EU Cyber Resilience Act and the German BSI fit. It is for vulnerability managers, PSIRT leads and product security engineers. For reading advisories by eye, see the How to Read a CVE Advisory Guide.

Key takeaways

  • CSAF 2.0 is the current OASIS Standard. It was approved on 18 November 2022, with one errata (January 2024) that only fixed the aggregator schema; CSAF 2.1 is still a committee specification draft.
  • Profiles decide what a document must contain. Security advisory, VEX, informational advisory and security incident response each make different fields mandatory on top of CSAF Base.
  • Product status and product identifiers are what make automation work. Without CPE or package URL identifiers in the product tree, a consumer cannot match an advisory to its estate.
  • Discovery is standardised. A provider-metadata.json at a well-known URL, a CSAF field in security.txt, signed files and ROLIE or directory feeds let consumers fetch advisories without scraping.
  • Consumers get most value first. Pulling CSAF feeds from your top suppliers into vulnerability management is a quick win even before you publish your own.

What CSAF is and how it relates to CVE, VEX and SBOM

CSAF is a document format and a distribution model for security advisories. It succeeds the older CVRF 1.2 format (the standard defines a conformance target for converting CVRF to CSAF). It sits alongside, not instead of, the other vulnerability data standards:

StandardAnswersRelationship to CSAF
CVEWhich vulnerability is this?CSAF advisories reference CVE IDs in /vulnerabilities[]/cve
CVSS, EPSS, KEVHow severe, how likely to be exploited, is it exploited?CVSS scores sit in the scores element; CSAF 2.1 drafts add EPSS and SSVC fields
SBOM (SPDX, CycloneDX)What components are in this product?Product identifiers (purl, CPE, hashes, SBOM URLs) let tools match CSAF advisories to SBOM components; the standard defines a "CSAF SBOM matching system" conformance target
VEXIs this product actually affected?VEX is a CSAF profile (csaf_vex); CycloneDX and OpenVEX are other VEX formats

See the VEX Guide, the SBOM Guide and the CVSS v4 Guide.

Anatomy of a CSAF document

A CSAF document is one JSON file with three top-level sections.

  • document: metadata. CSAF Base requires the category, CSAF version, publisher category, name and namespace, title, and tracking information: ID, initial and current release dates, revision history (date, number, summary), status (draft, interim or final) and version. It also carries the TLP distribution label, notes and references.
  • product_tree: every product the document mentions, organised in branches (vendor, product name, product version) with a product_id that the rest of the document refers to. The product_identification_helper must give at least one of: cpe, hashes, model_numbers, purl, sbom_urls, serial_numbers, skus or x_generic_uris.
  • vulnerabilities: one entry per vulnerability with its CVE or other IDs, CWE, notes, scores, threats, flags, product status and remediations.

Product status values

StatusMeaning in CSAF 2.0Typical consumer action
known_affectedKnown to be affected; action recommendedCreate remediation ticket
known_not_affectedKnown not to be affected; no remediation requiredSuppress scanner finding, keep the justification
under_investigationNot yet known; result to follow in a later releaseWatch; apply interim mitigations if exposure is high
fixed, first_fixedContain a fix, not necessarily the recommended oneConfirm deployed version
recommendedVendor-recommended fixed versionsTarget version for patching
first_affected, last_affectedBoundaries of the affected range in a release trainVersion range matching

Remediation categories

  • vendor_fix: an official fix from the original author, assumed to resolve the vulnerability fully unless noted.
  • workaround: a configuration or deployment change that avoids exposure.
  • mitigation: reduces risk but does not resolve the vulnerability.
  • no_fix_planned and none_available: explicit statements that there is no fix, which matter for end-of-life products and for risk acceptance decisions.

Remediations can also state restart_required, entitlements and a URL, which helps change planning. See the Remediation SLAs and Vulnerability Exceptions Guide.

A minimal excerpt

The illustrative excerpt below (fictional vendor, trimmed) shows how the parts connect: the product tree defines CSAFPID-0001 with a package URL, and the vulnerability entry refers to that ID in its product status and remediation.

"document": { "category": "csaf_security_advisory", "csaf_version": "2.0",
  "tracking": { "id": "EXAMPLE-SA-2026-0042", "status": "final", "version": "2" } },
"product_tree": { "branches": [ ... { "category": "product_version", "name": "4.2.1",
  "product": { "product_id": "CSAFPID-0001", "name": "Example Gateway 4.2.1",
  "product_identification_helper": { "purl": "pkg:generic/example/[email protected]" } } } ] },
"vulnerabilities": [ { "cve": "CVE-2026-00000",
  "product_status": { "known_affected": [ "CSAFPID-0001" ] },
  "remediations": [ { "category": "vendor_fix", "details": "Upgrade to 4.2.2",
    "product_ids": [ "CSAFPID-0001" ], "restart_required": { "category": "service" } } ] } ]

A consumer matches the purl against its inventory, sees known_affected, raises a ticket with the fixed version and schedules a service restart. Following the filename rule, this document is published as example-sa-2026-0042.json, and revision 2 replaces revision 1 under the same tracking ID.

The five profiles

Profile/document/categoryUse forExtra mandatory content
CSAF BaseAny value not reserved (for example "Example Company Security Notice")Catch-all; foundation for all othersDocument metadata and tracking only
Security incident responsecsaf_security_incident_responseResponding to a breach or incident, including one at another organisationA description, details, general or summary note, and an external reference
Informational advisorycsaf_informational_advisoryIssues that are not vulnerabilities, such as misconfiguration or hardening guidanceNotes and external reference; /vulnerabilities must not exist
Security advisorycsaf_security_advisoryVulnerabilities and remediations: the classic vendor advisoryProduct tree listing every referenced product, vulnerabilities with product status and remediation
VEXcsaf_vexStating that and why a product is or is not affectedCVE or IDs, notes, a product status list; an impact statement (flag or threat of category impact) for every not-affected product and an action statement for every affected one

VEX "not affected" flags are fixed values: component_not_present, vulnerable_code_not_present, vulnerable_code_cannot_be_controlled_by_adversary, vulnerable_code_not_in_execute_path and inline_mitigations_already_exist. Values of /document/category beginning with csaf_ are reserved for profiles in the standard, and validators must use the official profile when both an official and a private one exist.

Distribution and discovery

Section 7 of the standard defines 23 numbered requirements for publishing, grouped into roles that buyers can ask suppliers to meet.

RoleRequirementsWhat it means in practice
CSAF publisher1 to 4Valid documents, correct filenames, TLS, TLP:WHITE documents publicly accessible; only its own advisories
CSAF providerPublisher plus 5 to 7, at least one of 8 to 10, and either 11 to 14 (directory) or 15 to 17 (ROLIE)Access control for TLP:AMBER and TLP:RED, no redirects, a provider-metadata.json, and discoverability through security.txt, the well-known URL or DNS
CSAF trusted providerProvider plus 18 to 20Hashes for integrity, OpenPGP signatures and a published public key
CSAF lister6, 21 and 22Lists where issuing parties publish; does not mirror
CSAF aggregator1 to 6 and 21 to 23Mirrors signed documents from at least two disjoint issuing parties to give consumers one place to fetch
  • security.txt: a CSAF: field pointing to the provider-metadata.json, which must use https.
  • Well-known URL: /.well-known/csaf/provider-metadata.json under the issuer's main domain.
  • DNS path: csaf.data.security.domain.tld serving the provider metadata.
  • Directory distribution: one folder per year, plus index.txt and changes.csv so consumers can fetch only what changed.
  • ROLIE distribution: Atom-based feeds, service documents and category documents.
  • Filenames: derived from /document/tracking/id: lowercased, runs of characters outside a-z, 0-9, + and - replaced by one underscore, then .json appended.
  • Opt-outs: issuers set list_on_CSAF_aggregators or mirror_on_CSAF_aggregators to false if they do not want to be listed or mirrored.

Consuming CSAF as a vulnerability management team

  1. Inventory your suppliers' CSAF support: check security.txt and the well-known URL for your top 20 vendors by asset count and risk. Large vendors and agencies already publish: Red Hat moved to production CSAF on 1 February 2023 and stopped updating CVRF from 1 September 2023, and CISA publishes its IT and OT advisories, plus republished vendor CSAF files, in its CSAF repository on GitHub.
  2. Fetch with a downloader: open source tooling such as the gocsaf project's csaf_downloader retrieves documents from providers, and csaf_validator checks them against the schema. Verify signatures and hashes for trusted providers; reject unsigned documents claiming to come from one.
  3. Normalise product identifiers: map CPE and purl values to your CMDB and SBOM records. This is where most effort goes, and it pays back for scanners and SBOM tools too.
  4. Match and triage: raise tickets for known_affected matches, suppress scanner findings covered by a VEX known_not_affected statement (keeping the justification for audit), and track under_investigation items.
  5. Handle revisions: process updates by tracking ID and version; a later revision can move a product from under investigation to affected or change the recommended version.
  6. Enrich priority: combine CSAF data with KEV and EPSS before setting SLAs. See the Vulnerability Prioritisation Guide.
  7. Roll out gradually: start with one vendor in report-only mode and compare matches with manual triage for a month before letting CSAF drive tickets automatically.

Publishing CSAF as a product vendor

  1. Choose the target role: CSAF trusted provider is the role that lets customers automate safely, because documents are signed and discoverable.
  2. Get product identifiers right first: publish purl or CPE values that match what your SBOMs and customers' scanners see.
  3. Author with a validated tool: BSI published Secvisogram as a CSAF editor; any authoring tool should validate against the schema and the mandatory tests.
  4. Choose profiles deliberately: security advisories for fixes, VEX for "not affected" statements on high-profile CVEs in your components, informational advisories for hardening notices.
  5. Publish and advertise: provider-metadata.json, security.txt CSAF: field, well-known URL, signatures and public key; test with a provider checker such as gocsaf's csaf_checker, which tests against section 7.
  6. Integrate with your PSIRT process: CSAF generation should be part of advisory release, not an afterthought. See the Bug Bounty and Vulnerability Disclosure Guide.

Regulation: BSI and the Cyber Resilience Act

  • BSI: Germany's Federal Office for Information Security promotes CSAF 2.0 so that companies can compare advisories automatically with their inventory databases, publishes Technical Guideline TR-03191 on CSAF and points to open source provider, checker, aggregator, validator and CVRF converter tools.
  • EU Cyber Resilience Act: Regulation (EU) 2024/2847 entered into force on 10 December 2024. Manufacturers' reporting obligations for actively exploited vulnerabilities and severe incidents apply from 11 September 2026 (early warning within 24 hours, notification within 72 hours, final report within 14 days of a corrective measure for a vulnerability), and the main obligations apply from 11 December 2027. The Commission's summary pages do not name CSAF, and the recitals we reviewed only encourage machine-readable publication of security policies. CSAF is therefore a practical way to deliver security advisories to customers at scale, not a format the CRA prescribes. See the EU Cyber Resilience Act Vulnerability and Incident Reporting Guide.

CSAF 2.1: what is changing

As checked on 27 October 2026, CSAF 2.1 is at Committee Specification Draft 04 (21 October 2026), following drafts in May 2025, February 2026 and September 2026; it is not yet a Committee Specification or OASIS Standard. The current drafts add profiles for withdrawn and superseded advisories and a deprecated security advisory profile, and fields for EPSS and SSVC alongside CVSS. Draft details can change, so build on 2.0 now and plan to accept both versions.

Detection and measurement

  • Supplier coverage: share of critical suppliers that publish CSAF, and share that are trusted providers.
  • Automated match rate: share of CSAF advisories matched to assets without manual work.
  • Time from advisory publication to ticket for CSAF sources compared with manual sources.
  • False positive reduction: scanner findings closed by VEX known_not_affected statements.
  • For publishers: validation pass rate, time from advisory approval to CSAF publication and checker results.

Framework mapping

FrameworkReferenceRelevance
NIST SP 800-53 Rev. 5SI-5 Security Alerts, Advisories, and DirectivesReceiving and disseminating advisories
NIST SP 800-53 Rev. 5RA-5 Vulnerability Monitoring and Scanning, RA-5(2), SI-2 Flaw Remediation, SR-8 Notification AgreementsMatching, remediating and supplier notification
NIST CSF 2.0ID.RA-01, ID.RA-08, ID.AM-02, PR.PS-02Identifying vulnerabilities, disclosure handling, inventories, software maintenance
ISO/IEC 27001:2022 Annex AA.8.8 Management of technical vulnerabilitiesObtaining vulnerability information in a timely way
CIS Controls v8.1Control 7 Continuous Vulnerability ManagementVulnerability intake and remediation process
EU CRARegulation (EU) 2024/2847Vulnerability handling and reporting for manufacturers

Common failure patterns

  • Advisories without usable identifiers: product names only, no CPE or purl, so consumers fall back to manual matching and the format adds nothing.
  • Using CSAF Base when a profile applies: documents that should be security advisories miss mandatory fields and fail profile tests in consumer tools.
  • Unsigned feeds treated as trusted: consumers ingest documents without checking signatures, so a spoofed or tampered advisory could suppress real findings through a false "not affected" statement.
  • VEX without justification: "not affected" statements with no flag or impact statement, which auditors and customers cannot rely on.
  • Ignoring revisions: processing only the first release of an advisory and missing later changes in affected products or fixed versions.
  • Waiting for a mandate: treating CSAF as a compliance item rather than an automation opportunity, and so never consuming the feeds already available.

Practitioner checklist

  • List which critical suppliers publish CSAF and at which role.
  • Fetch CSAF with a downloader that verifies signatures and hashes.
  • Map CPE and purl identifiers to your CMDB and SBOM records.
  • Drive tickets from known_affected and suppress findings only with justified VEX statements.
  • Process every revision by tracking ID and version.
  • If you are a vendor, publish as a CSAF trusted provider with security.txt and the well-known URL.
  • Validate every document against the schema and profile tests before release.
  • Track CSAF 2.1 drafts and plan to support both versions.

Standards and references

Related NHI Mgmt Group resources: How to Read a CVE Advisory Guide · VEX Guide · SBOM Guide · Vulnerability Management Guide

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 27 October 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.