VSO announcements
VAP / VSO ·

VSO updates VAP references and research status; signature fix is on main only

This update summarizes merged changes since the 8 November website review. It separates current development, the frozen experimental distribution, and historical test records.

Signature fix: main only; fixed ZIP unchanged

VAP PR #68 strengthens Ed25519 input rejection in the experimental Python/Node interop demo. The fix is merged into main (35e56ba; included in e00ee5a). It is NOT included in the fixed vap-interop-demo-v0.2.0 ZIP, which remains at cbd4502. No replacement ZIP or new release is announced. Experimental extensions remain unadopted; independent third-party reproduction has not been established.

main snapshot · e00ee5a · Fixed prerelease ZIP · cbd4502 (not patched)

Current VAP source reference · 2026-11-12

The source remains v1.2.0 Draft 3 and includes editorial erratum E-2026-11-05 and the G-2 withdrawal. Its SHA-256 is unchanged from the main site’s 11 November reference; the VAP subsite previously displayed the older 3 November digest. This source check does not adopt a proposal. Experimental extensions remain unadopted; independent third-party reproduction is not established.

VAP v1.2.0 Draft 3 · e00ee5a · SHA-256 history

SHA-256: a35d5e082fd4b2f9921a78b1a71df263018320375e09fa18cf81533a18a19998

Bounded signature-verification correction — VAP #68

Both implementations explicitly reject small-order public keys and signature R points, noncanonical point encodings and scalars at or above the subgroup order before backend verification. The observed acceptance required the caller to substitute role trust keys; retaining the original trusted keys rejected the altered pack. Backend behavior differed by runtime. This is a local experimental input policy, not full prime-subgroup validation or a new VAP normative requirement. The recorded suite has 41 test methods (32 existing and 9 additional security methods); parameterized inputs are not a separate additive total. The 11 original vector/manifest files remain unchanged.

VAP documentation and change-process proposals — #66, #67, #69

Session/token validity is separated from permission intervals and revocation evidence. Source-monitoring notes distinguish reported proposals from adopted rules. Updated scope notes distinguish evidence verification from preventive control; LAP remains an external, unregistered proposal. Change-process and CHANGE-002 candidate inventories are unadopted drafts: formal identifiers, decisions and missing normative text/vectors remain unresolved. The IETF next-draft crosswalk is a proposal, not a submission or a published new revision. No public-review clock or committee approval is asserted.

VCP, SDK and historical reference records

VCP #30/#31 repair digest metadata and clarify bounded errata and ERASURE wording; they do not replace existing signed data or establish full VAP conformance. SDK #16/#17 distinguish development-document revisions from the existing release and reference an unadopted three-valued verification proposal. CAP and reference-implementation notes separate historical assessments from current source checks. Repository navigation, installation extras and CI maintenance have their own bounded scopes.

Explorer and public explanations

Explorer GUI #28 and API #30 distinguish reported metadata from verified anchors, precision and certification. The public GUI was already updated by site #216. This announcement does not reissue the Explorer or convert fixture endpoints into a native VCP v1.2 verifier. Existing product, operator, LAP relationship and incident-source corrections remain available; the PPC public-comment announcement remains a separate record.

First-party CI, not independent reproduction

At the 12 November 23:44 JST check, 36 workflow runs on the latest default-branch heads of the changed repositories had completed successfully. Three documentation repositories (vcp-sdk-spec, .github and cap-spec) had no run on their current head and are not counted as passed. The signature-fix CI and its source revision are linked separately. This article checks published CI records; it does not report a new independent execution or imply that each latest documentation commit reran every earlier experiment.

PR #68 · first-party signature-fix CI

Tests and CI are first-party synthetic checks. They do not establish adoption, certification, full conformance, production suitability or independent external reproduction. The frozen release retains its original 32-test/eight-pack record (two expected PASS and six expected FAIL per implementation); the later 41-method result belongs to the main fix, not that ZIP.

Merged source records

44 merged PRs · 8–11 November 2026
36 first-party workflow records · checked 12 November 23:44 JST

Historical evidence and distribution boundaries

Earlier announcements, including the Explorer record at 159 tests, remain unchanged. Website reference updates do not modify normative requirements, implementations, signed data, fixed distributions or K3. Open Draft PRs are excluded from merged outcomes.

2026-11-08 · VAP / VSO