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

PSIRT & disclosure

FIRST Traffic Light Protocol (TLP) 2.0, explained

Last updated:

TLP (Traffic Light Protocol) 2.0 is the current, FIRST-maintained standard for labeling how far a piece of potentially sensitive security information is allowed to travel once it leaves its source. Four labels — TLP:RED, TLP:AMBER (with an AMBER+STRICT variant), TLP:GREEN and TLP:CLEAR — tell a recipient, in plain language, who they may pass information on to. Published by FIRST's TLP Special Interest Group in August 2022 after a three-year, 50-plus-expert consultation, version 2.0 replaced the 2016-era TLP 1.0 and is now the baseline nearly every PSIRT, CSIRT and information-sharing community — including SoftWSP's own PSIRT-setup guidance — builds its sharing rules on top of.

Key takeaways

  • TLP (Traffic Light Protocol) 2.0 is the current, FIRST-maintained standard for labeling how far a piece of potentially sensitive security information is allowed to travel once it leaves its source.
  • TLP began as a UK government practice and became a FIRST standard in two steps, 2016 and 2022.
  • Each label sets who a recipient may pass information to, from named individuals only (RED) to anyone (CLEAR).
  • Two label changes are visible at a glance, and the rest is language and accessibility.

Where did TLP come from?

TLP began as a UK government practice and became a FIRST standard in two steps, 2016 and 2022. According to FIRST's own history in its version 2.0 release announcement, the protocol was "established in 1999 by the National Infrastructure Security Co-ordination Center (NISCC)" — a UK government body — to let analysts share threat information with industry without it leaking further than intended. It "gradually became well-known globally," and in 2015 "FIRST took the lead in unifying and standardizing the TLP." That work produced TLP 1.0, announced on 31 August 2016, which FIRST said had already "achieved 'de facto' international standard status" among CSIRTs, trust communities, ISAOs, government agencies and researchers.

In 2019, over 50 incident-response professionals reconvened as the FIRST TLP Special Interest Group (TLP-SIG) to improve "content, language, supporting materials, and accessibility." TLP 2.0 was released on 5 August 2022; FIRST states it is "authoritative from August 2022 onwards," and its download page marks TLP 1.0 "Deprecated August 2022."

What do the four labels mean?

Each label sets who a recipient may pass information to, from named individuals only (RED) to anyone (CLEAR). TLP's §3 defines each label with a one-line meaning and guidance for sources and recipients:

Label Meaning (verbatim) Who it's for
TLP:RED "For the eyes and ears of individual recipients only, no further disclosure." Information that "cannot be effectively acted upon without significant risk for the privacy, reputation, or operations of the organizations involved." In a meeting, RED is limited to those present.
TLP:AMBER "Limited disclosure, recipients can only spread this on a need-to-know basis within their organization and its clients." Information needing support to act on, but risky if shared outside the organizations involved. "Clients" are, by TLP's own definition, people or entities receiving cybersecurity services from the organization — included by default so they can act to protect themselves.
TLP:AMBER+STRICT A restriction on AMBER: "if the source wants to restrict sharing to the organization only, they must specify TLP:AMBER+STRICT." Same risk profile as AMBER, but the source does not want the organization's clients included.
TLP:GREEN "Limited disclosure, recipients can spread this within their community." Information useful for wider community awareness, shareable with "peers and partner organizations... but not via publicly accessible channels." Default community, if undefined, is "the cybersecurity/defense community."
TLP:CLEAR "Recipients can spread this to the world, there is no limit on disclosure." Information with "minimal or no foreseeable risk of misuse," shareable "without restriction," subject to ordinary copyright rules.

Two label-writing rules apply across the board, per §1.b: labels "MUST not contain spaces and SHOULD be in capitals," and they "MUST remain in their original form, even when used in other languages" — a document can be translated, but TLP:AMBER stays TLP:AMBER.

What changed from TLP 1.0?

Two label changes are visible at a glance, and the rest is language and accessibility. FIRST's release announcement lists:

  1. TLP:WHITE became TLP:CLEAR— same meaning, different word.
  2. TLP:AMBER+STRICT was added, to restrict AMBER-level information to "the recipient's organization only," closing a gap where AMBER's default client inclusion was not always what the source wanted.

FIRST also "Removed synonyms and colloquialisms to improve accessibility for non-native English speakers and ease of translation," added formal Community, Organization and Clients definitions, and added an RGB/CMYK/Hex colour table. CISA set a hard cutover date in its own fact sheet, stating it "will officially move from TLP 1.0 to TLP 2.0 on November 1, 2022," while FIRST hoped for full use worldwide "by January 2023."

How is TLP applied in messages, documents and feeds?

TLP 2.0's §2 sets explicit MUST/SHOULD rules, using the RFC-2119 convention the standard cites, for three channels:

  • Messaging (email, chat): the label "MUST" appear "directly prior to the information itself," and "SHOULD" be in the email subject line; where the labelled content does not run to the end, mark where it ends.
  • Documents: the label "MUST" appear "in the header and footer of each page," "SHOULD" be at least 12-point type "for users with low vision," and right-justification "is recommended."
  • Automated information exchanges (e.g. MISP, IEP): the mechanism is deliberately left open — "this is left to the designers of such exchanges, but MUST be in accordance with this standard."

The colour table sets label text in #FF2B2B (RED), #FFC000 (AMBER), #33FF00 (GREEN) or #FFFFFF (CLEAR) on a black #000000 background, and says sources "SHOULD adhere to the TLP color-coding to ensure enough color contrast" for readers with low vision. For verbal discussions, CISA's fact sheet says "participants should assume information is TLP:CLEAR if the speaker does not provide a designation," and FIRST's use cases (2A) likewise tell chat rooms and presentations to announce a default label.

How does a PSIRT use TLP in coordinated disclosure?

A device maker's PSIRT typically narrows or widens the label as a case moves from private report to public advisory. TLP itself does not prescribe these stages; the table applies FIRST's definitions to a common coordinated vulnerability disclosure flow.

Stage Typical label Why, in TLP's own terms
Embargoed technical details sent to a named engineer at a chip or module supplier TLP:RED or TLP:AMBER+STRICT RED limits it to the individual; AMBER+STRICT lets the supplier's own team work on it without passing it to the supplier's other customers
Advance notice to OEM and integrator customers before a fix ships TLP:AMBER AMBER includes the recipient's clients, so integrators can warn the operators they serve
Coordination with a national CSIRT TLP:AMBER or TLP:AMBER+STRICT Use case 6B lets a national CSIRT share AMBER with constituents and partners; 6A forbids that for AMBER+STRICT without permission
Indicators or detection rules for an ISAC or sector group TLP:GREEN with the community named Use case 1C: GREEN plus accompanying instructions defining the community
Published advisory, CSAF file and release notes TLP:CLEAR "no limit on disclosure"

The source sets the label, so the PSIRT should record which label each outgoing message carried and why — useful later when regulators or customers ask how a case was handled, for example under the CRA's reporting duties. The PSIRT services framework places this inside FIRST's wider PSIRT model.

What do FIRST's TLP use cases add?

FIRST's TLP Use Cases page answers the edge cases the four definitions leave open. The ones that matter most for a vendor PSIRT:

Use case Question FIRST's answer
1B Can AMBER go to a recipient's service providers? Only with instructions; FIRST recommends allowing it only "as TLP:AMBER+STRICT"
1E One document, mixed sensitivity? Label the whole with the most restrictive label and mark portions, e.g. a GREEN block of indicators inside AMBER+STRICT
3A Notifying an affected organisation directly, no restrictions needed? TLP is optional
4A / 4B Recipient outsources services — can it share AMBER or AMBER+STRICT? "No, not without permission"
5C / 5D Can AMBER received via an ISAC go to its members? Only if the originator sent it to the group, or to you as a group member
7A Recipient wants wider sharing than the label allows? Ask the originator; set up a standing agreement

How does TLP appear in a CSAF advisory?

In CSAF, TLP lives in document.distribution.tlp, with a label and an optional url pointing at the TLP definition used. The published CSAF 2.0 standard still uses the TLP 1.0 vocabulary: valid labels are AMBER, GREEN, RED and WHITE, and the default url is FIRST's TLP page. Its optional test 6.2.10 warns when the label is missing, requirement 7.1.4 says a TLP:WHITE document "MUST be freely accessible," and 7.1.5 says TLP:AMBER and TLP:RED documents "MUST be access protected," served from a different path with automatable authentication. A minimal synthetic CSAF 2.0 fragment for a public advisory:

"distribution": {
  "text": "Disclosure is not limited",
  "tlp": { "label": "WHITE", "url": "https://www.first.org/tlp/" }
}

CISA's own files follow this pattern: the CISA CSAF repository keeps public ICS advisories in a white folder, and advisory ICSA-26-316-04 (Siemens RUGGEDCOM APE1808) carries "label": "WHITE".

The CSAF 2.1 draft — Committee Specification Draft 04, dated 21 October 2026 — moves to TLP 2.0:

CSAF 2.0 (OASIS Standard) CSAF 2.1 (CSD04 draft)
Label values AMBER, GREEN, RED, WHITE AMBER, AMBER+STRICT, CLEAR, GREEN, RED
Default label none defined CLEAR
Public-file rule TLP:WHITE freely accessible TLP:CLEAR freely accessible
Access-protected AMBER, RED AMBER, AMBER+STRICT, RED
Closed-group field — sharing_group (UUID), e.g. for multi-party CVD

The draft notes that "+STRICT" is technically an extension of AMBER but is offered as its own label value to simplify parsing, and its 2.0-to-2.1 conversion table maps WHITE to CLEAR and keeps the other three labels. For the wider CSAF picture, see CSAF/VEX for device makers.

What does a labelled disclosure look like? A worked example

Synthetic example, for illustration only. A researcher reports a command-injection flaw in the web UI of a fictional router whose Wi-Fi module comes from an outside supplier.

  1. Day 0:the PSIRT acknowledges the report and sends the module supplier's named contact the proof of concept marked TLP:AMBER+STRICT, with an instruction that the supplier may involve its incident-response contractor (use case 1B).
  2. Day 20:a fix is built; three ISP customers who ship the router get an advance note marked TLP:AMBER, so they can prepare firmware rollouts for their subscribers.
  3. Day 25:indicators for detecting exploitation go to a sector ISAC as TLP:GREEN, limited in writing to ISAC members.

What is TLP explicitly not?

TLP is a sharing convention, not a legal control. §1.c says: "TLP is not a formal classification scheme. TLP was not designed to handle licensing terms, nor information handling or encryption rules. TLP labels and their definitions are not intended to have any effect on freedom of information or 'sunshine' laws in any jurisdiction." CISA's fact sheet puts it plainly: "Unlike formal classification systems, TLP is not legally binding." §1.e adds that TLP "is distinct from the Chatham House Rule, but may be used in conjunction when appropriate."

Responsibility sits with the source: §1.f–h say the source "is responsible for ensuring that recipients of TLP-labeled information understand and can follow TLP sharing guidance," may add restrictions recipients "must" follow, and a recipient wanting wider sharing "must obtain explicit permission from the source." No TLP body adjudicates violations; FIRST calls TLP "optimized for ease of adoption, human readability and person-to-person sharing."

What are the common TLP mistakes?

  • Using retired or home-made labels. TLP 2.0 has four labels plus the AMBER+STRICT restriction, and "Only labels listed in this standard are considered valid by FIRST." TLP:WHITE was retired in 2022 and survives only as a CSAF 2.0 enum value.
  • Forgetting that AMBER includes clients. If an advisory should stop at the recipient's own staff, label it TLP:AMBER+STRICT.
  • Putting the label only in a cover note. Subject line for email, header and footer for documents, so a forwarded body still shows it.
  • Treating TLP as an NDA. Pair it with a contractual control where confidentiality must be enforceable.
  • Assuming two systems agree on "TLP 2.0 support." Automated exchange is undefined by the standard; check how each schema, such as CSAF 2.0 versus 2.1, encodes labels.

Most CRA, NIS2 and PSIRT vulnerability-handling obligations assume a working information-sharing discipline underneath them — see the PSIRT setup guide for the surrounding machinery. On the SoftWSP platform, CISA ICS advisories are ingested only from the CISA repository's public white folder; restricted TLP material is not read or produced. SoftWSP's PSIRT setup work builds TLP-aware intake and advisory-sharing workflows into a device maker's disclosure program: see PSIRT services.

This guide quotes FIRST's own published TLP 2.0 standard text (Introduction, Usage, and Definitions sections), FIRST's TLP Use Cases page, FIRST's official TLP 2.0 and TLP 1.0 release announcements, the OASIS CSAF 2.0 standard and CSAF 2.1 CSD04 draft, and CISA's September 2022 TLP 2.0 fact sheet; it is not a substitute for the full standard at first.org/tlp, and does not constitute legal advice. FIRST's TLP 2.0 standard document is copyrighted material (© 2015—2026 Forum of Incident Response and Security Teams, Inc.) — the short excerpts quoted here are reproduced for commentary and identification, not as a substitute for reading the source.

Put this guide to work

Frequently asked questions

What is the Traffic Light Protocol, and who maintains it?

The Traffic Light Protocol (TLP) is a labeling system, standardized and maintained by the Forum of Incident Response and Security Teams (FIRST), for marking how far a piece of potentially sensitive information may be shared. Per the standard's own Introduction: "The Traffic Light Protocol (TLP) was created to facilitate greater sharing of potentially sensitive information and more effective collaboration. Information sharing happens from an information source, towards one or more recipients. TLP is a set of four labels used to indicate the sharing boundaries to be applied by the recipients. Only labels listed in this standard are considered valid by FIRST." The current version — "FIRST Standards Definitions and Usage Guidance — Version 2.0" — states it "is the current version of TLP standardized by FIRST. It is authoritative from August 2022 onwards." TLP itself predates FIRST's standardization: per FIRST's own version-2.0 release announcement, it was "established in 1999 by the National Infrastructure Security Co-ordination Center (NISCC)" in the UK, and "FIRST took the lead in unifying and standardizing the TLP in 2015," publishing TLP 1.0 on 31 August 2016 before TLP 2.0 superseded it.

What are the four current TLP labels, and what does each one mean?

Verbatim from the standard's §3: **TLP:RED** = "For the eyes and ears of individual recipients only, no further disclosure" — sources use it when information "cannot be effectively acted upon without significant risk for the privacy, reputation, or operations of the organizations involved," and recipients "may therefore not share TLP:RED information with anyone else." **TLP:AMBER** = "Limited disclosure, recipients can only spread this on a need-to-know basis within their organization and its clients" — usable when information "requires support to be effectively acted upon, yet carries risk to privacy, reputation, or operations if shared outside of the organizations involved." A note to AMBER adds: "if the source wants to restrict sharing to the organization only, they must specify TLP:AMBER+STRICT." **TLP:GREEN** = "Limited disclosure, recipients can spread this within their community" — shareable with "peers and partner organizations within their community, but not via publicly accessible channels," and "may not be shared outside of the community" (where community defaults to "the cybersecurity/defense community" if undefined). **TLP:CLEAR** = "Recipients can spread this to the world, there is no limit on disclosure," usable when information "carries minimal or no foreseeable risk of misuse," and shareable "without restriction" subject to standard copyright rules. The standard also defines **Community**, **Organization** and **Clients** as terms of art: Clients are "those people or entities that receive cybersecurity services from an organization" and are "by default included in TLP:AMBER so that the recipients may share information further downstream in order for clients to take action to protect themselves," with the definition extending to "stakeholders and constituents" for teams with national responsibility.

What actually changed between TLP 1.0 (2016) and TLP 2.0 (2022)?

Per FIRST's own 5 August 2022 release announcement, "the new version of the TLP results from a thorough consultation with over 50 security industry experts over three years" — the FIRST TLP Special Interest Group (TLP-SIG), re-convened in 2019 — "with the goals to standardize, unify and modernize the content and language and provide improved supporting materials." FIRST itself lists "the most significant changes" as: "Removed synonyms and colloquialisms to improve accessibility for non-native English speakers and ease of translation"; "Focused on consistent language and terminology, adding definitions for community, organization, and clients"; "Added a colors table to include RGB, CMYK, and hexadecimal color codes"; "TLP:WHITE has become TLP:CLEAR"; and "Added the TLP:AMBER+Strict label to highlight information that is restricted to the recipient's organization only." FIRST TLP-SIG co-chair Don Stikvoort (Open CSIRT Foundation) is quoted: "We are increasingly spreading more confidential and sensitive information inside our community... We need systems that are easy to use, simple to understand, and straightforward enough that translation does not impact the meaning." FIRST targeted full worldwide adoption of 2.0 by January 2023; CISA's own fact sheet confirms it "will officially move from TLP 1.0 to TLP 2.0 on November 1, 2022."

How does TLP actually get applied — in email, in documents, and in automated feeds?

The standard's §2 sets MUST/SHOULD rules (defined per RFC-2119) for three channels. **Messaging** (email, chat): "TLP-labeled messaging MUST indicate the TLP label of the information, as well as any additional restrictions, directly prior to the information itself. The TLP label SHOULD be in the subject line of email." **Documents**: "TLP-labeled documents MUST indicate the TLP label of the information, as well as any additional restrictions, in the header and footer of each page. The TLP label SHOULD be in 12-point type or greater for users with low vision," and right-justifying the label "is recommended." **Automated information exchanges** (e.g. MISP, IEP): the standard explicitly does not define the mechanism — "TLP usage in automated information exchanges is not defined: this is left to the designers of such exchanges, but MUST be in accordance with this standard" — while §1.d confirms TLP "may be used in automated information exchange systems." CISA's own TLP 2.0 fact sheet adds a verbal-discussion convention not spelled out in the FIRST text: in speeches or meetings, "participants should assume information is TLP:CLEAR if the speaker does not provide a designation."

Is TLP a legal classification system, or does it override freedom-of-information law?

No, on both counts, and the standard says so directly. §1.c: "TLP is not a formal classification scheme. TLP was not designed to handle licensing terms, nor information handling or encryption rules. TLP labels and their definitions are not intended to have any effect on freedom of information or 'sunshine' laws in any jurisdiction." CISA's fact sheet echoes this for a general audience: "Unlike formal classification systems, TLP is not legally binding." §1.e also distinguishes TLP from the Chatham House Rule — "TLP is distinct from the Chatham House Rule, but may be used in conjunction when appropriate" — and §1.f-h place responsibility squarely on the source: the source must ensure recipients understand the sharing guidance, may add further restrictions recipients must follow, and a recipient needing to share more widely than a label allows "must obtain explicit permission from the source."

Sources

  1. FIRST: Traffic Light Protocol (TLP) — FIRST Standards Definitions and Usage Guidance, Version 2.0 (official standard page: Introduction, Usage §2, Definitions §3, label-by-label text, RFC-2119 note; authoritative from August 2022)
  2. FIRST: "FIRST Releases Traffic Light Protocol Version 2.0 with important updates" — official press release, 5 August 2022 (history: NISCC 1999, FIRST unification 2015, TLP-SIG consultation, list of significant changes, Don Stikvoort quote)
  3. FIRST: "FIRST announces Traffic Light Protocol (TLP) version 1.0" — official press release, 31 August 2016
  4. CISA: Traffic Light Protocol — Moving to Version 2.0 (fact sheet, September 2022: official 1 November 2022 CISA adoption date, verbal-discussion default-to-CLEAR convention)
  5. FIRST: TLP Use Cases (seven groups of operational use cases clarifying TLP 2.0: service providers, mixed labels, notifications, ISACs, national CSIRTs)
  6. OASIS: Common Security Advisory Framework Version 2.0, OASIS Standard (§3.2.1.5.2 TLP label enum, §6.2.10, §7.1.4–7.1.5)
  7. OASIS: Common Security Advisory Framework Version 2.1, Committee Specification Draft 04 (21 October 2026) — TLP 2.0 labels, default CLEAR, sharing groups
  8. CISA CSAF repository on GitHub: public (TLP:WHITE) ICS advisory folder
Was this guide helpful?

In this cluster

FIRST Traffic Light Protocol (TLP) 2.0, explained