VSO publishes VCP research checks and updates implementation status
VeritasChain Standards Organization (VSO) has incorporated research proposals and experiments on VCP hash inputs, signature coverage, numeric representation and evidence-preserving migration. The website now distinguishes the published specification, first-party pre-release code and unadopted research, with links to the reviewed GitHub records.
Research publication — unadopted
The experimental extensions and proposed repairs remain unadopted and not effective. Independent third-party reproduction has not been established. Merging research material or passing internal CI does not establish native VCP/VAP conformance, production suitability or normative adoption.
What is available, and what each item means
| Item | Current status and limits |
|---|---|
| VCP v1.2 specification | Published specification. Known integrity, signature, schema and numeric issues remain under review. The research does not silently repair or replace the current normative text or schema. |
| vcp-sdk v0.1.0 | A first-party, pre-release offline core in Python, TypeScript and MQL5, available from GitHub. It is not published on PyPI or npm, not an implementation of vcp-sdk-spec, and not an independent implementation. Its interpretations are library choices, not adopted specification changes; VALID is not a conformance claim. |
| VCP research candidate and checker | A closed synthetic subset for field placement, exact units, ISO/ns consistency and retained source-to-derived evidence. It reads supplied evidence copies; it does not migrate real VCP records or verify the complete native format. |
| VAP three-valued verification proposal | The latest document clarification preserves a known failure even when a separate required check is unknown. This remains a proposal, not a change to the current normative algorithms. |
Internal results for the reviewed candidate
PR #16 was merged on 26 October 2026 at 3f99ba3. The reviewed HEAD 7ca7855 passed Python 3.12 and 3.14 CI. The saved internal runs record:
| Item | Current status and limits |
|---|---|
| Synthetic evidence bundles | 62 expected outcomes matched: 6 PASS, 43 FAIL, 13 UNASSESSED |
| Exact-unit cases | 14 expected outcomes matched: 5 PASS, 4 FAIL, 5 UNASSESSED |
| Regression tests | 39 passed, with no skips or expected failures |
| Saved-report controls | Full-report recomputation matched; all 8 deliberately altered reports were rejected |
These counts overlap and are not an independent-validation total. Negative and unsupported outcomes are intentional: matching an expected FAIL or UNASSESSED is not successful verification of that evidence. The final review fixed malformed-policy crashes and cases where unknown inputs hid known failures. Both policy roles now return structured FAIL for ten non-object inputs through the CLI.
Decisions and evidence still outstanding
The complete native schema, field presence and identifiers, other units/time bases, inner approval signatures, production authority and migration operations remain open. Coverage accounts for 74 numeric inventory entries: nine selected exercises, one claimed-time scalar with different semantics and 64 unassessed entries. Neither test keys nor synthetic receipts establish an independent signer or external time proof.
Existing records and the earlier kit remain separate
The current specification and schema, existing implementations, previously signed data, fixed distributions and K3 were not changed by this research PR. The 24 October VAP reproduction kit (kit b1af0dc; target c8ccf97) remains a separate fixed experiment, including its Linux/Docker requirements. The new read-only Python experiment neither replaces that ZIP nor upgrades its historical results.
Website corrections
English, Japanese and Chinese pages have been aligned on profile status, first-party SDK availability, demo limitations, CAB approval and independence, and the current individual Internet-Draft revisions. Dated posts remain historical records. VCP full conformance to VAP remains unestablished; Internet-Draft publication is not IETF endorsement or Working Group adoption.