Two engineers at a desk reviewing code together on a large monitor

PSIRT & disclosure

The FIRST PSIRT Services Framework, explained

Last updated:

FIRST's PSIRT Services Framework is a free, 91-page reference document — currently Version 1.1, dated Spring 2020 and marked TLP:WHITE (shareable without restriction) — that describes the services a Product Security Incident Response Team (PSIRT) can provide. Published by FIRST (the Forum of Incident Response and Security Teams), it is not a certifiable standard: it is a planning and training catalogue that most device makers draw on when they design, not audit, their vulnerability-handling team.

Key takeaways

  • FIRST publishes v1.1 as a PDF, as an HTML page, and in Arabic, Spanish, French, Russian and Chinese translations; v1.0 stays online for reference.
  • The framework is explicit that a PSIRT needs "the autonomy to maintain an independent and objective position" and that it should "report to an executive" with the authority to back its findings.
  • Reachability (Function 2.1.1).
  • The framework itself has no levels.

What actually is the framework?

FIRST describes its Services Frameworks (PSIRT and the parallel CSIRT one) as "high level documents detailing possible services that computer incident response teams (CSIRTs) and product incident response teams (PSIRTs) may provide," developed "by recognized experts from the FIRST community" and driven by FIRST's Education Advisory Board. Its stated purpose is to "provide a foundation for the development of new training material," though FIRST notes it is "used in a much wider scope, for example when defining an initial service catalogue for new teams." A notice on page 2 is explicit about its status: the document describes what FIRST "believes are best practices," offered "for informational purposes only," with FIRST disclaiming liability for its use. Unlike ISO/IEC 29147 or 30111, there is no certification or conformity scheme attached to it.

FIRST publishes v1.1 as a PDF, as an HTML page, and in Arabic, Spanish, French, Russian and Chinese translations; v1.0 stays online for reference. The introduction defines a PSIRT as the entity that focuses on "the identification, assessment and disposition of the risks associated with security vulnerabilities within the products" an organisation produces or sells, and insists it is "not an independently operating group" but part of the secure development lifecycle.

How is the framework structured?

The framework organises everything into four nested levels:

Level Definition, per the framework Role
Service Area Groups "services related to a common aspect," to "facilitate understanding" Top-level category
Service "A set of recognizable, coherent actions towards a specific result on behalf of or for the constituency of an incident response team" The unit most teams plan around
Function "An activity or set of activities aimed at fulfilling the purpose of a particular service"; a function "might be shared and used in the context of several services" Reusable building block
Sub-function A lower-level activity performed as part of a function Implementation detail

Each Service and Function is specified with a "Description" field and a "Purpose and Outcome" field — a consistent template throughout the document, which is one reason it gets reused as a gap-analysis checklist even though it was written as training material.

What has to exist before the services start?

Ahead of the six numbered Service Areas, an "Operational Foundations" section lists what an organisation needs "to plan, establish, and effectively operate a PSIRT":

Group Components Typical evidence
I. Strategic Executive Sponsorship, Stakeholder, PSIRT Charter, Organizational Model, Management and Stakeholder Support Named executive sponsor, stakeholder list, charter with mission and services
II. Tactical Budget, Staff, Resources and Tools Funding model, roles and skills, intake channel, secure communications, tracking system
III. Operational Policies and Procedures, Evaluation and Improvements Written procedures, metrics, post-incident lessons learned

The framework is explicit that a PSIRT needs "the autonomy to maintain an independent and objective position" and that it should "report to an executive" with the authority to back its findings.

Which organizational model fits a device maker?

The framework describes three models and says plainly that there is "no single one-size fits all" template:

Model How it works Fits Main challenge (per the framework)
Distributed A small core PSIRT sets policy and takes intake; product teams' security representatives triage and fix Large, diverse product portfolios The people who triage and ship fixes do not report to the PSIRT
Centralized A larger PSIRT staff, reporting to one or more senior executives, monitors sources, triages and even provides code fixes "A smaller organization and/or an organization with a homogenous product portfolio" Cost of a specialised team that scales poorly as the portfolio grows
Hybrid Mixes both, chosen by corporate structure, portfolio size and diversity, and development strategy Growing portfolios, acquisitions Defining who owns which decision

For a device maker with one firmware platform shared across product lines, the centralized description usually matches; once each product line has its own firmware tree and release train, the distributed or hybrid models become more realistic.

What are the six Service Areas?

Service Area Services (numbering from the framework)
1 — Stakeholder Ecosystem Management 1.1 Internal Stakeholder Management; 1.2 Finder Community Engagement; 1.3 Community and Organizational Engagement; 1.4 Downstream Stakeholder Management; 1.5 Incident Communications Coordination within the Organization; 1.6 Reward Finders with Recognition & Acknowledgement; 1.7 Stakeholder Metrics
2 — Vulnerability Discovery 2.1 Intake of Vulnerability Reporting; 2.2 Identify Unreported Vulnerabilities; 2.3 Monitoring for Product Component Vulnerabilities; 2.4 Identifying New Vulnerabilities; 2.5 Vulnerability Discovery Metrics
3 — Vulnerability Triage and Analysis 3.1 Vulnerability Qualification; 3.2 Established Finders; 3.3 Vulnerability Reproduction
4 — Remediation 4.1 Remedy Release Management Plan; 4.2 Remediation; 4.3 Incident Handling; 4.4 Vulnerability Release Metrics. Purpose: "Highlight the processes and mechanisms required to release and announce a remedy to stakeholders and downstream vendors"
5 — Vulnerability Disclosure 5.1 Notification; 5.2 Coordination; 5.3 Disclosure; 5.4 Vulnerability Metrics. Purpose: "Provide transparency to stakeholders and partners through collaboration with finders, coordinators and downstream vendors to responsibly disclose vulnerabilities and fixes"
6 — Training and Education 6.1 Training the PSIRT; 6.2 Training the Development Team; 6.3 Training the Validation Team; 6.4 Continuing Education for all Stakeholders; 6.5 Provide Feedback Mechanisms

What do the services ask of a hardware or firmware maker?

Several functions read like a device maker's to-do list:

  • Reachability (Function 2.1.1). Publish PSIRT contact details in product documentation and on the website, register them in CSIRT/PSIRT lists, communicate them to CVE Numbering Authorities, reserve addresses such as psirt@ or security@, and accept encrypted reports (S/MIME, PGP or an HTTPS form). A security.txt file (RFC 9116) is a common way to publish the contact in machine-readable form. Sub-function 2.1.2.2 adds "process reports in isolation," because a PSIRT is an obvious target for malicious attachments.
  • Component monitoring (Service 2.3). The framework splits vulnerabilities into a product's own code, vendor-internal components and third-party components, open source included. Function 2.3.1 asks for an inventory whose outcome is a "completed bill of materials for all products"; 2.3.3 points to the NVD where suppliers publish no advisories; 2.3.5 asks for automated notification of the affected development teams. For firmware, that inventory is a firmware SBOM.
  • Remedy release plan (Service 4.1). Define which products and versions are supported and when support ends (4.1.1), how a remedy is delivered — "hot fix, patch, maintenance releases, firmware updates" — and whether it deploys remotely, by the customer or on site (4.1.2), plus a release cadence with documented out-of-band exceptions (4.1.3).
  • Security advisories (Function 5.3.2). A template with "title, summary, CVE(s), supported product impact and status, acknowledgement, references and revision history," a defined CVE assignment process, and publication in a machine-readable format such as CSAF. See CSAF and VEX for device makers for the format work.

How do the services line up with FIRST's maturity levels?

The framework itself has no levels. FIRST's separate PSIRT Maturity Document groups the same services into Level 1 (Basic), Level 2 (Intermediate) and Level 3 (Advanced). Its Annex 7 sample checklist places them like this:

Service Area Level 1 Level 2 Level 3
Foundations Executive sponsorship, stakeholders, budget, policies Charter, organizational model, product registry with dependency mapping, baseline metrics Organizational policies, cost of vulnerability management
1 Stakeholders — Internal and downstream stakeholders, incident communications Direct finder engagement, community membership, stakeholder metrics
2 Discovery Intake; PSIRT e-mail address with PGP key Discovering unreported vulnerabilities Monitoring product component vulnerabilities, discovery metrics
3 Triage Intake process and qualification workflows Repeat finders, reproduction capability —
4 Remediation Analyse options, document the fix Formal remedy management plan Advanced incident handling, release metrics
5 Disclosure CVE/CVSS, templates, stakeholder communication, finder acknowledgement Notification system, vulnerability metrics Industry coordination playbook
6 Training — PSIRT training, feedback mechanisms Development team training, continuing education

The FIRST PSIRT maturity model guide walks through each level and its mapping to EU Cyber Resilience Act duties.

Which framework terms matter in a vulnerability-handling policy?

The framework's definitions fix a few terms that are easy to use loosely:

  • Remediation (or Remedy) — "a change made to a product or online service to remove or mitigate a vulnerability," typically a binary replacement, configuration change, or source patch and recompile; "mitigations are also called workarounds or countermeasures."
  • Coordinator — an "optional participant that can assist vendors and finders in handling and disclosing vulnerability information."
  • Bug Bar — "criteria that defines the types of bugs that qualify as a security vulnerability," used so reports that meet the bar are processed through the PSIRT's standard operating procedure.
  • Embargo — "a hold on the publication of vulnerability details until affected vendors are able to release security updates or mitigations and workarounds."
  • Advisory — an "announcement or bulletin that serves to inform, advise, and warn about a vulnerability of a product."

How does the framework differ from ISO/IEC 29147 and 30111?

FIRST PSIRT Services Framework ISO/IEC 29147 / 30111
Nature Voluntary best-practice/training document, TLP:WHITE, free Internationally standardised processes (ISO/IEC, paywalled full text)
Certifiable No No formal certification scheme either, but widely cited as the reference process standard
Best used for Designing the team, its stakeholder relationships, and its full service catalogue Defining the specific vulnerability-disclosure and handling process steps a policy commits to
Legal standing None cited; FIRST disclaims liability for use Referenced informally by regulators and buyers as good practice, but not itself an EU harmonised standard either

Most hardware and firmware makers use both together: ISO/IEC 29147/30111 or the Cyber Resilience Act's Annex I Part II text to write the disclosure policy commitments, and the FIRST framework to design the team, stakeholder engagement and metrics around it. The CRA's Article 14 reporting duties apply from 11 September 2026. See how to set up a PSIRT for a hardware company, the CVD process under ISO/IEC 29147 and 30111 and CRA 24h/72h vulnerability reporting for the operational and legal detail.

What does a first-year service catalogue look like for a small device maker?

A worked example, using a synthetic company: a 40-person maker of Linux-based industrial gateways with three product lines on one shared firmware platform, two security-minded engineers and no PSIRT.

  1. Foundations firstThe CTO signs as executive sponsor; a one-page charter names the mission, the products in scope and the services offered; the budget covers part-time time from the two engineers plus tooling.
  2. Pick the modelOne firmware platform and a small team point to the centralized model: the two engineers triage and coordinate fixes for all three lines.
  3. Run the Level 1 services2.1 intake (security@, PGP key, security.txt), 3.1 qualification with a written bug bar, 4.2 remediation, 5.3 advisories with CVE IDs and CVSS scores.
  4. Add component monitoring earlyBuild an SBOM per firmware release (Function 2.3.1) and watch for vulnerabilities in it (2.3.3), because most gateway CVEs arrive through open-source components rather than finder reports.
  5. Write the remedy release planSupported versions and end-of-support dates (4.1.1), signed firmware updates as the delivery method (4.1.2), a quarterly cadence with an out-of-band path for critical issues (4.1.3).
  6. Defer what can waitFinder rewards (1.6), proactive testing (2.4) and stakeholder-wide continuing education (6.4) move to year two.
  7. Measure from month oneTime to acknowledge, time to fix and advisories published feed the Evaluation and Improvements component.

The result is a catalogue of about ten services that two people can sustain, mapped to numbers anyone familiar with the framework will recognise.

What are the common mistakes when adopting the framework?

  • Treating all 28 services as mandatory instead of picking the ones the charter commits to.
  • Skipping the Operational Foundations, then finding that no executive will approve an out-of-band firmware release.
  • Running intake without component monitoring, so a library CVE that affects every product is noticed only when a customer asks.
  • Writing advisories without a fixed template, which makes later CSAF publication harder.
  • Citing the framework as proof of compliance; it carries no regulatory weight.

Where does SoftWSP fit against the framework?

SoftWSP covers part of Service 2.3: it extracts a firmware SBOM as CycloneDX 1.6, matches components against NVD, CISA KEV, FIRST EPSS and CISA CSAF ICS advisories, sends watchlist alerts on a CVE ID or vendor/product, and drafts a cited report for an engineer to review. Triage, remediation and disclosure decisions stay with the manufacturer. See PSIRT service.

This guide summarises FIRST's freely published PSIRT Services Framework v1.1 (TLP:WHITE) and PSIRT Maturity Document, read on 2026-11-12; it is general information, not legal advice.

Put this guide to work

Frequently asked questions

What is the FIRST PSIRT Services Framework?

It is a high-level reference document from FIRST (the Forum of Incident Response and Security Teams) describing the services a Product Security Incident Response Team (PSIRT) can provide. Version 1.1, marked TLP:WHITE and dated Spring 2020, states its own purpose as providing "a foundation for the development of new training material" that in practice is "used in a much wider scope, for example when defining an initial service catalogue for new teams." It is guidance, not a certifiable standard: FIRST's own notice says the document describes what FIRST "believes are best practices" for "informational purposes only."

How is the framework structured?

Four levels: Service Areas group related services; a Service is "a set of recognizable, coherent actions towards a specific result"; a Function is an activity that fulfils part of a service's purpose and can be shared across services; and a Sub-function is a lower-level activity within a function. Before the six numbered Service Areas, a separate foundational section covers Strategic, Tactical and Operational building blocks (such as executive sponsorship, charter, and evaluation) that an organisation needs before running any of the services.

What are the six Service Areas?

FIRST's v1.1 text names them as Service Area 1, Stakeholder Ecosystem Management; Service Area 2, Vulnerability Discovery; Service Area 3, Vulnerability Triage and Analysis; Service Area 4, Remediation (remedy release management plan, remediation, incident handling and release metrics); Service Area 5, Vulnerability Disclosure (notification, coordination, disclosure and vulnerability metrics); and Service Area 6, Training and Education (training the PSIRT, development and validation teams, continuing education for all stakeholders, and feedback mechanisms).

How does the FIRST PSIRT Services Framework relate to ISO/IEC 29147 and ISO/IEC 30111?

They cover the same problem from different angles. ISO/IEC 29147:2018 and ISO/IEC 30111:2019 are internationally standardised processes for, respectively, receiving and disclosing vulnerability reports and handling them internally. FIRST's PSIRT Services Framework is a broader, non-certifiable catalogue of the services, organisational models and capabilities a PSIRT can build, intended as training and planning material rather than an auditable process standard. A hardware or firmware maker typically uses ISO/IEC 29147/30111 (or the CRA's own Annex I Part II text) to define its disclosure obligations, and the FIRST framework to design the team and service catalogue that carries them out.

Does following the framework satisfy the EU Cyber Resilience Act's PSIRT-related duties?

Not directly. The framework is FIRST's own voluntary guidance; no EU Official Journal citation was found making it a harmonised standard for the Cyber Resilience Act, and FIRST's own notice disclaims liability for its use. It is, however, a practical way to plan the team and processes that then have to meet the CRA's actual legal duties in Annex I Part II and Article 14. See the CRA vulnerability-reporting guide below for the legal deadlines themselves.

Sources

  1. FIRST: Product Security Incident Response Team (PSIRT) Services Framework, Version 1.1 (PDF, Spring 2020, TLP:WHITE)
  2. FIRST: PSIRT Services Framework v1.1 — standards page
  3. FIRST: Services Framework overview
  4. ISO/IEC 29147:2018 — Vulnerability disclosure (ISO catalogue)
  5. ISO/IEC 30111:2019 — Vulnerability handling processes (ISO catalogue)
  6. FIRST: PSIRT Maturity Document (Levels 1-3, Annex 7 sample checklist)
  7. OASIS: Common Security Advisory Framework (CSAF) Version 2.0
  8. RFC 9116: A File Format to Aid in Security Vulnerability Disclosure (security.txt)
  9. Regulation (EU) 2024/2847 (Cyber Resilience Act), Official Journal

About this guide

SoftWSP Firmware Security Desk

Covers the PSIRT & disclosure cluster: 15 guides.

    Was this guide helpful?

    All guides in this cluster: PSIRT & disclosure

    Bug bounty programs for IoT and hardware vendors, explainedHow a paid bug bounty program differs from a free vulnerability disclosure policy, what CISA's own federal directive says that distinction is, and what real IoT/hardware program scope, safe-harbor and payout terms look like — including a documented case where unclear scope turned a bounty into a legal dispute.CISA CIRCIA: critical infrastructure incident reporting, explainedWhat the Cyber Incident Reporting for Critical Infrastructure Act (CIRCIA) requires: who counts as a covered entity, what a substantial cyber incident is, the proposed 72-hour/24-hour deadlines, and why none of it is mandatory yet as of November 2026.CRA vulnerability reporting: the 24-hour early warning, 72-hour notification and final reportHow CRA Article 14 reporting works: 24-hour early warning, 72-hour notification, final report, the ENISA Single Reporting Platform and what to prepare.ISO/IEC 29147 and 30111: the coordinated vulnerability disclosure process explainedHow ISO/IEC 29147 (disclosure) and ISO/IEC 30111 (handling) define coordinated vulnerability disclosure, and how CISA and FIRST apply the same process.CVE Numbering Authority (CNA) rules, explained: how CVE IDs actually get assignedHow the CVE Program's CNA Operational Rules v4.2.1 actually govern CVE ID assignment — the 72-hour first-refusal clock, 24/72-hour publication deadlines, Scope Definitions, and what a CVE Record must contain — straight from the Program's own rules text.ENISA's European Vulnerability Database (EUVD), and what NIS2 Article 12 requiresWhat the EU Vulnerability Database (EUVD) actually is: its NIS2 Article 12 legal basis, the critical/EU-CSIRT-coordinated/exploited categories it publishes, how it relates to CVE/NVD/CISA KEV, and ENISA's own CNA role.The FIRST CSIRT Services Framework, explainedWhat FIRST's Computer Security Incident Response Team (CSIRT) Services Framework actually defines: its five Service Areas, the CSIRT-vs-PSIRT distinction, Enterprise/Coordinating/National team types, and how it pairs with the parallel PSIRT Services Framework.FIRST PSIRT Maturity Document: the three levels explainedFIRST's PSIRT Maturity Document: what its Basic, Intermediate and Advanced levels each require, the service areas it scores, and how it pairs with the PSIRT Services Framework v1.1.FIRST Traffic Light Protocol (TLP) 2.0, explainedWhat FIRST's TLP 2.0 standard actually says: the four labels (RED, AMBER, AMBER+STRICT, GREEN, CLEAR) in the source's own words, the rules for using TLP in messages, documents and automated exchanges, and what changed from TLP 1.0.NIST SP 800-216 (federal vulnerability disclosure guidelines), explainedNIST SP 800-216 (May 2023) — the Federal Coordination Body (FCB) and Vulnerability Disclosure Program Office (VDPO) framework that implements the IoT Cybersecurity Improvement Act's disclosure mandate, built on ISO/IEC 29147 and 30111.The NVD CVE backlog: what it means for your PSIRT, and what Vulnrichment fills inNIST says its own NVD enrichment backlog is still growing faster than it can clear it. What that means for a device maker's PSIRT, what CISA's Vulnrichment project adds in the meantime, and why the fix NIST describes is automation, not AI, today.PSIRT response SLA examples: published deadlines from regulators, coordinators and vendorsPSIRT response deadlines from the CRA, BOD 26-04, CERT/CC, Project Zero, ZDI, FDA and vendor policies, each sourced and dated, and how to set your own SLA.How to set up a PSIRT and vulnerability disclosure process for a hardware product companyStep-by-step: set up a PSIRT and vulnerability disclosure process for a device or hardware company, aligned with ISO/IEC 29147, ISO/IEC 30111, FIRST and the EU Cyber Resilience Act.Vulnerability disclosure policy template and security.txt guideA copy-ready disclosure policy template for device makers, checked against CISA's VDP template and ISO/IEC 29147, plus security.txt (RFC 9116).

    The FIRST PSIRT Services Framework, explained