I debated in my original proposal suggesting that we update the CA/B F requirements to say that we have to specify the versions of the documents we comply with. This seems like it would align better with the proposal from Microsoft for CAs to update the CP/CPS after BR updates. I think it also fits better with how WebTrust audits work. This has some side effects we'd want to walk through more.
On Wednesday, February 18, 2026 at 3:03:08 PM UTC-8 Ryan Hurst wrote: > Jim, > > I may not have been clear in my original ask, so let me restate it. I > wasn’t suggesting audits are “based on the CPS.” I understand WTBR is the > audit criteria and it maps to specific BR versions. My question was about > the bridge step: when you move from a WTBR criterion to evaluating what a > CA actually does, what role does CP/CPS specificity play? > > Your walkthrough answers that directly. You described the workflow as WTBR > criterion > mapped BR section/version > CP/CPS > evidence/testing. That > makes the CP/CPS the document auditors use to connect the criterion to the > CA’s implementation choices and the evidence trail, which is exactly why > the specificity question matters. > > This is also where I think Trev’s Proposal 1 runs into a structural > problem. If the CP/CPS is primarily incorporation-by-reference (“we comply > with the BRs”), then in the bridging step you described the auditor is sent > back to the same BR text they already have open rather than to CA-specific, > testable commitments. Two points you raised make that especially > problematic. > > First, version lag. You noted WTBR criteria often map to older BR > versions. A CP/CPS that mostly points to “the BRs” without describing the > CA’s actual current operational practice and parameters makes that > translation harder, not easier. > > Second, multi-framework use. You noted the same CP/CPS is used across > multiple audit frameworks. If it is too thin to describe a CA’s actual > practices in auditable terms, it fails multiple audiences at once. > > This is not abstract. Bug 2009525 is a concrete example. Language like > “may be backdated by a few hours” without a bound or threshold is not > something an auditor can evaluate as a measurable commitment under the > bridge process you described. And it does not meet what Apple, Mozilla, and > Chrome have now said explicitly they expect from a CP/CPS. > > So your process description ends up supporting the opposite conclusion > from Proposal 1: if the CP/CPS is the bridge from WTBR criteria to actual > practice, it needs enough specificity to describe the CA’s implementation > choices in a way that can be mapped to evidence and tested. > > Ryan > > On Wed, Feb 18, 2026 at 4:25 AM 'Jim Scardelis' via CCADB Public < > [email protected]> wrote: > >> Since Ryan asked for auditors’ input .…. >> >> >> >> As a WebTrust practitioner, the most current CA/B Forum Baseline >> Requirements are not what we audit against. Neither is the CPS. In my >> personal opinion (and only my opinion, not that of my employer or anyone >> else), both proposals would make it much more difficult to conduct the >> various audits that CAs subject to the CA/B Forum BRs have to go through >> (see the WebTrust Engagement Applicability Matrix >> <https://www.cpacanada.ca/-/media/site/operational/ms-member-services/docs/webtrust/engagement-applicability-matrix_final_aoda-compliant.pdf?rev=d0091ed381f74635ba2736c16978ec78&hash=1FD5E0964D6C3E689353DEAE5A21E2B2> >> >> for details on those, but it’s at least three, all of which use the same >> CPS.) >> >> More details on my reasoning below, but the [tl;dr] is IMHO, the current >> language is already optimized, and keeps the CPS readable. >> >> >> >> For the purposes of this discussion, we would be auditing against the >> WebTrust >> Principles and Criteria for Certification Authorities – TLS Baseline >> <https://www.cpacanada.ca/-/media/site/operational/ep-education-pld/docs/01618-ms-24-3852_webtrust---tls-baseline-v29.pdf?rev=e0678ebc91614665b1b2d0a25a50d9bf&hash=5837E2C482F5162302AA0A61A1EE6B38> >> >> framework (“WTBR”) . WebTrust frameworks aren’t updated as often as the >> CA/B Forum updates their requirements, and there is typically a delay in >> time before we have to use a new version - that is determined by the >> starting date of the engagement period. >> >> >> >> So, for example, a client undergoing their annual WTBR audit that had >> their previous audit period end on January 31, 2025, would now be audited >> for the period from February 1, 2025 to January 31, 2026 – against WebTrust >> Principles and Criteria for Certification Authorities – SSL Baseline >> Requirements v2.8, since the version linked above (v2.9) is required for >> engagement periods commencing on or after April 1, 2025. >> >> >> >> This matters because the each of the WTBR standards is based on a >> particular version of the CA/B Forum Baseline Requirements: >> >> - WTBR v2.8 is based on CA/B Forum Baseline Requirements version 2.0.1 >> - WTBR v2.9 is based on CA/B Forum Baseline Requirements version 2.1.2 >> - WTBR v2.10, which is for engagement periods commencing on or after >> December 1, 2025, is based on CA/B Forum Baseline Requirements version >> 2.1.7 >> >> >> >> When conducting the audit, we consider the CA’s compliance with each of >> the applicable WebTrust criteria. For WTBR, most of the criteria include >> references to one or more specific sections of the CA/B Forum BRs. As >> stated in the WTBR standards, “The practitioner is directed to consider the >> referenced Section(s) as part of assessing the CA’s compliance with each >> criterion.” The referenced section(s) mentioned refer to section in* the >> version of the CA/B Forum BRs that the WebTrust standard was based on, *not >> the version that CAs are currently required to conform to*.* >> >> >> >> Additionally, for a CA that would be subject to the CA/B Forum TLS BRs, >> the WTBR audit is one of at least three that we will be auditing them >> against using the same CPS document – WebTrust for CAs (required for all of >> types of CAs), WTBR and WebTrust for CAs – Network Security. The CPS is >> where their practices, policies and procedures that are applicable to all >> of them are disclosed, and each of the WebTrust standards has requirements >> around the CP and CPS. >> >> >> >> As a result, each section of a CA’s CPS is topical and not necessarily >> directly related to the CA/B Forum BRs. Every CA I’ve audited includes a >> reference to conformance with the latest version of the Baseline >> Requirements .… issued by the CA/Browser Forum because WTBR Principle 1, >> Criterion 1 includes >> >> “The CA discloses on its website .… its commitment to conform to the >> latest version of the Baseline Requirements .… issued by the CA/Browser >> Forum.” This criterion in each of the above versions, for example, also >> references section 2.2 and 3.2.6 of the CA/B Forum BRs, so in deciding if a >> client being audited under WTBR v2.8 is compliant with this criterion, I >> would look at their CPS to see if it is posted on their website; are >> cross-certificates disclosed; and is there a statement either on their >> website that discloses their commitment to conform to the latest version of >> the BRs. And then, I would open up the CA/B Form BRs version 2.0.1, look >> at sections 2.2 and 3.2.6 and consider whether they had met those as well. >> >> The only time that I would consider more recent versions of the CA/B >> Forum BRs would be if the CAs practices differed from what the referenced >> sections said in a way that was less strict. In that case, I can look at >> the more current versions of the BRs in effect during the audit period say, >> and for that one criterion, consider if their practice was because of the >> changed requirement. >> >> >> >> Finally, WebTrust for Certification Authorities v. 2.2.2, which covers >> topics like the process for verifying various elements that are provided by >> an applicant for a certificate, groups criteria differently than the CA/B >> Forum BRs. For example, in WebTrust for CAs, there is one criterion and >> illustrative control that covers all types of authenticated certificates, >> while the CA/B Forum BRs spread them out by type, with repeated >> requirements for parts of each type. >> >> >> >> All of which means that including the text of CA/B Forum BRs throughout >> the CPS would actually make auditors’ jobs much more difficult. >> >> >> >> Jim >> >> >> >> >> *Jim Scardelis *M.S., CISSP, CISA, CEH, CCSK, QSA, PCI Secure Software, >> Secure SLC, P2PE, & P2PE Application Assessor, PCIP, CIPP/US, CIPP/C, >> CIPP/E, CIPT, CTT+ >> >> Technical Lead - Crypto/Digital Trust >> 1.866.254.0000 ext. 584 <(866)%20254-0000> >> >> >> >> SOC • PCI • HITRUST • FedRAMP • ISO >> >> >> >> >> <https://outlook.office.com/bookwithme/user/[email protected]?anonymous&ep=signature> >> >> *Book time to meet with me >> <https://outlook.office.com/bookwithme/user/[email protected]?anonymous&ep=signature>* >> >> >> >> *Upcoming Out of Office dates**:* >> >> *March 1-8* >> >> *March 19-20* >> >> *April 2-3* >> >> *March 19-20* >> >> *April 2-3* >> >> >> >> *From:* [email protected] <[email protected]> *On Behalf Of *Ryan Hurst >> *Sent:* Friday, February 13, 2026 15:23 >> *To:* Trevoli Ponds-White <[email protected]> >> *Cc:* CCADB Public <[email protected]> >> *Subject:* Re: CP/CPS Content Consistency Proposal >> >> >> >> Thanks Trev, I appreciate you engaging directly on this. >> >> I agree that duplicating required BR text often doesn't add clarity and >> just creates maintenance headaches. Your example around reasons 6-16 is a >> good one. That's the kind of discretionary parameter that belongs in the >> CPS, in my opinion. Same logic applies to validation methods and profile >> constraints where the obligation is mandatory but operational choices vary. >> Same with EKU. The option space has narrowed, but the CA should still state >> which values appear in which profiles. >> >> Where I think there's more to work through is what belongs in a CPS when >> a CA incorporates by reference. If "only describe optional or discretionary >> items" becomes the principle, we lose something. A requirement can be >> mandatory in the BRs while the way a CA implements it still involves >> choices that affect risk posture. The distinction is less "required versus >> optional" and more "normative obligation versus implementation choice." >> Incorporation by reference establishes the obligation. It doesn't always >> make the CA's posture visible. >> >> One concern with the "SHOULD NOT include details about required >> practices" framing. It could discourage CAs who want to be more >> transparent. We should be careful not to create a norm that penalizes >> specificity. Taken to its conclusion, that rewards minimal disclosure. CAs >> that say less create less surface area for findings. >> >> This also has audit implications. Auditors audit against the CPS. If the >> CPS just says "we comply with the BRs," what is the auditor actually >> testing against? The CPS says we comply, the auditor confirms we said we >> comply. Without implementation specifics in the document, there's less to >> measure. Over the last several years we've seen more CAs default to "we >> comply with RFC 5280" without specifying profiles. In my experience, few >> auditors in this space have the technical depth to independently identify >> which implementation details to drill into. The CPS is their roadmap. If >> it's vague, the audit will be too. >> >> On validation methods, the CPS doesn't need to restate what each BR >> method is. But it should state which methods the CA uses, which it doesn't, >> and any constraints on how they're implemented. That's not duplicating the >> BRs. It's disclosing choices. >> >> We saw why this matters with the TLS-ALPN-01 flaws in early 2022. When >> vulnerabilities turned up in a CA's implementation of 3.2.2.4.20, the >> ecosystem needed to quickly assess who else was exposed. That triage only >> works when CAs enumerate what they use. If every CPS just says "we use >> approved methods," root programs have no starting point and the response >> turns into a polling exercise. >> >> Bug 1962829 is worth looking at too. Microsoft's CPS had a specific >> commitment about keyEncipherment in RSA subscriber certs. Turned out to be >> a typo that didn't match practice. A third-party researcher caught it. It >> was just a typo, and not a consequential one but it was only catchable >> because the commitment was specific and measurable. If we remove >> specificity what happens when the error actually matters? >> >> The alternative to a detectable error is an undetectable one, even if >> minor. If a CA can't detect its own implementation errors, removing that >> information from the CPS means no one else can either. That raises a fair >> question about what else might be going unnoticed. >> >> On your threshold question, if someone has to leave the CPS and interpret >> external documents to understand what the CA actually does, the document >> isn't doing its job. That doesn't mean restating every clause. It means the >> CA's choices, bounds, and constraints should be findable in one place. >> >> I wrote up some broader thoughts on this at >> https://us01.z.antigena.com/l/FG0A7qEmnyPiHyI0eCOu8gQjvFUTAT4M4ubAcFCh-muvY0hB-kkaWfJcwKMMLLaFeoWZULx_LW3ZzkuznRdow8KxmeMJHNSt-QRWMPCErCOnXFi3LXIvkY5QMtggHUrvYmBnpYx-UfK7l5T7uaeAgXIz6pXR92hiI0uJ0m087UxE5PSMacZ34Wzm-8rVJV6oyy >> >> for anyone interested in the longer version. >> >> Ryan Hurst >> >> >> >> On Fri, Feb 13, 2026 at 2:20 PM 'Trevoli Ponds-White' via CCADB Public < >> [email protected]> wrote: >> >> Thanks Ryan, you are definitely a good voice to weigh in on this topic. >> Looking at some of your examples I think they are definitely a mix of what >> I would consider BR copy pasta vs unique information. For example I think >> the 24 hour revocation SLA for reasons 1-5 is very clear and not optional. >> I’m not sure what it adds for all CAs to repeat this? Alternatively for >> reasons 6-16 it’s more interesting because there is an optional time frame >> there. Under my first proposal I would expect a CP/CPS to state the >> timeline a CA tries to use. Extended Key Usage is another good example >> where we have evolved the BRs to a place where there are less options than >> there used to be. Rather than CAs listing items that are not allowed or >> strictly required a section like this one “7.1.2.2.5 Cross-Certified >> Subordinate CA Extended Key Usage – Restricted” has more value if it >> specifically addresses if “Any other value” is present and why. >> >> I think you also bring up something that I think is fundamental to >> resolve. What is the threshold we think is fine for parties to have to >> reference the various baseline requirements to understand what a CA is >> doing? I think this is core to the discussion. The BRs allow incorporation >> by reference because of this it’s a very common practice. Section 3 is a >> good example of this. Most CAs do not describe the validation methods. Do >> we think that CAs should add the descriptions into their docs so that >> consumers don’t have to look at the BRs to understand them? >> >> >> >> On Friday, February 13, 2026 at 1:42:40 PM UTC-8 Ryan Hurst wrote: >> >> Thanks, Trev. I care a lot about this topic and think it’s an important >> discussion for the ecosystem. >> >> I put some thoughts on the bug >> (https://us01.z.antigena.com/l/FPW1RbEmpif0KE8eTNvI9ou9J5_-B9fKsAM4bxiDdqUCU4kqT7eXWXH-I_wj4J0-ppm72j0D_86-ay0VwYsrHhOSvDMkFL-Orauu_rMQ4aTIKpxxodw82x2VTB2l8qKX_XFpD7EPXjyKCacm8PDCe8XukRqnQcKMYFAVmLzuIJAXq8xD0DjQG4K-BfBWuq0NjidrXA5AYLj9im >> >> ) that frame how I’m thinking about CP/CPS structure and governance. I >> initially tried to reply here but realized I had to request access, so the >> bug seemed like the easiest place to respond in the moment. >> >> Happy to discuss any of those points here and hear other perspectives. >> >> I’d be especially curious to hear from auditors about how they are >> evaluating these documents in practice. How much specificity do they rely >> on in the CP/CPS itself versus interpreting incorporation by reference? >> That seems like a critical input to getting this right. >> >> Looking forward to the discussion. >> >> Ryan Hurst >> >> >> >> On Friday, February 13, 2026 at 11:53:40 AM UTC-8 Trevoli Ponds-White >> wrote: >> >> Hello! >> >> We want to start a conversation about bringing more clarity and >> consistency to expectations for CP/CPS content. We posted this here so that >> non-CA/B F members can provide feedback and because we can think of several >> ways to do it. An update to the Baseline Requirements, CCADB policy update, >> or an update to one of the Root programs’ policies. >> >> This isn’t the first incident about CP/CPS content the community has had >> but it’s a good place to start with for background >> https://us01.z.antigena.com/l/FPW1RbEmpif0KE8eTNvI9ou9J5_-B9fKsAM4bxiDdqUCU4kqT7eXWXH-I_wj4J0-ppm72j0D_86-ay0VwYsrHhOSvDMkFL-Orauu_rMQ4aTIKpxxodw82x2VTB2l8qKX_XFpD7EPXjyKCacm8PDCe8XukRqnQcKMYFAVmLzuIJAXq8xD0DjQG4K-BfBWuq0NjidrXA5AYLj9im >> >> . >> >> We have two proposals to start the discussion. If people have other >> proposals or feedback about these, please share them. >> >> Proposal 1 – Update Section 2.2 of the Baseline Requirements >> >> Change the existing language: >> >> *The CA SHALL publicly give effect to these Requirements and represent >> that it will adhere to the latest published version. The CA MAY fulfill >> this requirement by incorporating these Requirements directly into its >> Certificate Policy and/or Certification Practice Statements or by >> incorporating them by reference using a clause such as the following (which >> MUST include a link to the official version of these Requirements):* *[Name >> of CA] conforms to the current version of the Baseline Requirements for the >> Issuance and Management of Publicly-Trusted TLS Server Certificates >> published at **https://www.cabforum.org* <https://www.cabforum.org/>*. >> In the event of any inconsistency between this document and those >> Requirements, those Requirements take precedence over this document.* >> >> >> >> To something like: >> >> >> >> *The CA MUST publicly give effect to these Requirements and represent >> that it will adhere to the latest published version. The CA MAY fulfill >> this requirement by incorporating these Requirements by reference into >> its Certificate Policy and/or Certification Practice Statements. If the CA >> does this the reference must be in Section 1.1 and MUST list which >> documents it is referencing by title and must include the link to the >> document’s landing page i.e. “Baseline Requirements for the Issuance and >> Management of Publicly**‐Trusted TLS Server Certificates >> **https://us01.z.antigena.com/l/FUQCybnjgy1X-0YVScoqasPJWJb3V-3CD7gOvK~BOBXSh2DXJMt6mfskHLIqXZaSFWzQIbRhQ0blOTJQ4fCCBEZeKGXw5TqelWhsY-Zana54sbV8GIGgypqe6-xnKvrUcPT77fqVPaBf7ZJAQm1vjs2ZTv_XDacJHQKH9Etk06MQ1yEixr9Pg85Di6-flh3huEGc7j >> >> r/baseline-requirements/* >> <https://us01.z.antigena.com/l/joTNwbdm8ssK0CdpqZ0SMk6k3~M8sVmUpXNQ1S9GtMMSSFSzY2WkW2-giA0OjhEakgb-OInefTmeD5jn2bMSvsgp2ll13M2d4J~le91BIUauyIasXFjz9eaeraRmUhr4n9F7vwdxWkojt3HrTY4-wXN~gwdcdjgCQr5ZwCEOCzs2Eofmn5FbRO2yQLvLhKHfQl4FoLMr9~i~MMT5Ri4HAE~Ydggl>*”. >> >> The CA MAY also include the following in Section 1.1: “In the event of any >> inconsistency between this document and those Requirements, those >> Requirements take precedence over this document.*”* CAs that incorporate >> requirements by reference SHOULD NOT include details about required >> practices. CAs that incorporate requirements by reference MUST include >> details in places where the referenced documents express optional >> requirements i.e. “No stipulation, SHOULD, SHOULD NOT, MAY, optional.* >> >> >> >> In practice this would result in, for example, Section 4.2.1 (for TLS >> certs) CAs would have to describe if they reuse validation data but not >> required items such as: “Applicant information MUST include, but not be >> limited to, at least one Fully-Qualified Domain Name”. >> >> >> >> Proposal 2 – Adopt a style similar to Matter PKI by the Connectivity >> Standards Alliance (CSA) >> >> When CA's submit a CPS to the CSA it is laid out where the requirement is >> stated. Following that there is a spot for the CA response when required. >> Example: >> >> *4.2 - Certificate Application Processing* >> *It is the responsibility of the CA/RA to verify that the information in >> a Certificate Application is accurate.* >> *CA Response: <insert response>* >> >> *CA Response:* >> >> *Before issuing Matter Certificates, the CA verifies Vendor identity & >> information through the following process: 1) Vendor requests a Distributed >> Compliance Ledger (DCL) challenge to prove ownership of their DCL private >> key; 2) The CA verifies the DCL challenge; 3) The CA retrieves Requester >> data for the RAD and Certificate Application; 4) The CA verifies the >> Vendor's information against a known database; 5) The CA generates the RAD >> and Certificate Application with the verified data; 6) Vendor signs the RAD >> and Certificate Application; and 7) The CA creates the Vendor for PAI >> issuance.* >> >> Given that we have been moving the CA/B F requirements in a direction >> where we have greater specificity and less leeway we think this is a good >> time to revisit more specific requirements for CP/CPS content as well. >> >> Thanks! >> >> Trevoli Ponds-White >> >> Amazon Trust Services >> >> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "CCADB Public" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion visit >> https://us01.z.antigena.com/l/DSx1PkghwwODeZZfljhkaKXMonK8vcIUp0GqfOBa0qYD4ME9ohSHFonX-JckjlyGkL76iWb5L7uWcESG3W3hWu-aETvv58sdrRph7uiBJgxmPmMqgno-~q41KZnEMg53BkiQaEd5~V-ZZXM8T8weGCvL3kC_vArvG0I88i4sVr3eMckWUmT1LSeV2_xo~-XDzwKoDZV3gQm9Bl~OIOQhxJG4jGvLDqF1Foh-RHyXmFYDnXi5_ktadilI >> >> >> <https://us01.z.antigena.com/l/ap6gj6MYAL9ZwMm7VQXYIiczsem83EiayYGM_lXnwZoKBCLUj6mBURx~cUxZfKzXxzDgTnXxNvPNkN77m1G4MThi6djhkyQGsHRjfT_lIbzmkfWhCA9KptlMDxGsF_GNMOkljqags3G8l4GW1IFEJ2_FGl~yt0MhbTtD2w8h0tUWKef0sQHSHij-6gMh1_QCEE2NnLY_H_kqfPNehLox-trfadsEl1T-ArWZMFzPppve~9PEPyMW-Ez~21Do1wRWejVjoFSvV4W4sqUyl2~VoBL> >> . >> >> -- >> You received this message because you are subscribed to the Google Groups >> "CCADB Public" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion visit google.com >> <https://us01.z.antigena.com/l/5i2KwrDeqf4mM4WJlKcAXLqqKhDIhTPCciAnwbkYQ4TlJI8H6j2Z-0srSrGldCM7Gf1~3ojJVLS7gIH4uZT2GvSSlFJOi~UaBXSZFEanElXOMM90WWtvN2catkB2llVS~vKb~zflMK~Ft~10hfzvnILtrtqIjtAKzJ07JZGSnOM4xn97Hl6vxTpBRhiJB1hm6q~RZ1x4oCXb7snkySoKO6YJb6klIOATUOwfavBy8KK-QdiWObwux67f3wJWUHnb36alufZcwbE3136asPfFOwbl8CoBFI8gTpUXowmPo2Aid~Qy5x8mrgTNe-RivCtWRe> >> . >> >> -- >> You received this message because you are subscribed to the Google Groups >> "CCADB Public" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> > To view this discussion visit >> https://groups.google.com/a/ccadb.org/d/msgid/public/CH3PR14MB6770F0AFA92068203185D9A8FE6AA%40CH3PR14MB6770.namprd14.prod.outlook.com >> >> <https://groups.google.com/a/ccadb.org/d/msgid/public/CH3PR14MB6770F0AFA92068203185D9A8FE6AA%40CH3PR14MB6770.namprd14.prod.outlook.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "CCADB Public" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/a/ccadb.org/d/msgid/public/0e6b5571-e53b-44d7-b626-6ebb0d3e072bn%40ccadb.org.
