Verifiable Agent Trust Envelope

VATE v0.4.1 Discussion-Draft Pre-Release Notes v0.4.1 is a maintenance release of the verifier-side admission and receipt discussion draft. It corrects the AP2 replay case's authoritative inputs and collects reproducible partial implementation records and coverage guidance. Changes Since v0.4.0 AP2 replay input correction (#55). deny-ap2-hnp-replay now supplies the normalized payment-authority artifact and VATE-local consumed-state context through sut_inputs. The case binds both input digests and checks the supplied replay state. This fixes the fixture's input contract; it does not implement a durable or atomic replay store or redefine AP2 payment semantics. Partial native implementation records (#56, #58, #59). The source archive includes Vaara stale-status admission and Bolyra / x402 request-binding records, their pinned inputs, native output, projections, comparison reports, and replay instructions. Coverage and responsibility guides (#60). The coverage guide separates native execution, reached predicates, reported checks, and corpus comparison. The EVC, ACS, and ACLE-MCP comparison describes admission, enforcement, attenuation, replay, and outcome-evidence responsibilities at pinned source revisions. It defines no new adapter or compatibility profile. Dependency maintenance. The package-private TypeScript workspace uses Vitest 4.1.11 with its updated lockfile. No new direct dependency is introduced. Fixed Corpus and Compatibility Semantic profile: VATE-AL2-Verifier-Admission-v0.3 Machine-readable artifact line: 2026-09 Cases / manifest artifacts: 76 / 217 Corpus digest: sha-256:02da4ca9257547872ecbcce9728b8bb256afd2f77676d5daef89fd6c831b2bb0 The profile, schemas, and artifact-line identifiers are unchanged. The added replay-state fixture and corrected case change the corpus bytes: v0.4.0 contains 76 cases / 216 artifacts with digest b2a281e372b2e1d6b49be219c715fa69c0b2be237d29a6e1f0dda9c0659b6130. An adapter for the replay case must supply both authoritative input references. Pin the tag and corpus digest when preparing or comparing a result. Recorded Execution Scope The Vaara and Bolyra records evaluate the fixed v0.4.0 source cc072ef86f54791213a3e603a65b2f24f64b1b6d and its 216-artifact corpus. Each exercises one partial case; the other 75 cases were not executed. Their native refusals agree with the selected VATE decisions, while missing checks, reason differences, and, for Bolyra, a missing legacy receipt-fixture reference leave both selected cases failing VATE comparison. Bolyra's two additional controls are not additional VATE cases. The Pulse record remains tied to source 5a37f87de0190da44e619b1800261637e83dd7ed, 75 cases / 212 artifacts, and digest 988aae7d03dd5bb743e8e03e6ab1120ce8735a4837ac818ffd9d665de0c1e370. It records three executed cases: two matches and an amount-overrun mismatch (deny versus VATE attenuate), plus 72 explicit skips. Including these records in v0.4.1 preserves their evidence; no new native Vaara, Bolyra, or Pulse execution against the v0.4.1 corpus is reported. See the run register for the complete pins and the separate AlgoVoi adapter-result slices. Validation and Citation The technical release gate records the repository, conformance, schema, TypeScript, and source-archive checks. run checks the repository fixtures; compare checks a submitted result against one corpus; verify-bundle checks the local report digest chain. These checks do not by themselves establish independent native execution. Cite 10.5281/zenodo.22680789 for the archived v0.4.1 GitHub tag. The existing concept DOI identifies the release series. The exact v0.4.0 DOI 10.5281/zenodo.22218860 continues to identify that historical snapshot. VATE remains an Apache-2.0 discussion draft, not production-ready. The bounded implementation records and passing repository checks do not imply certification, endorsement, production approval, or general compatibility. Acknowledgements Thanks to Vaara (Henri Sirkkavaara) for publishing the native verification scripts and results used in the one-case external-SUT evaluation. Thanks to Viswanadha Pratap Kondoju (@saneGuy, Bolyra) for publishing the runner, wrapper and native results used in the Bolyra/x402 one-case partial external-SUT evaluation.