[stormer78]

Moves @openvtc/vti-tsp-js to the released TSP 1.0 specification (tswg-tsp-specification#86, ccda6b3), in step with affinidi-tsp 0.3.0 (affinidi/affinidi-tdk-rs#967, published) and the reference tsp_sdk 0.16.0.

The change

• Packs YTSP-BAA (1.0), was YTSP-AAC (Rev 3). unpack, decodeEnvelope and peekRevision refuse every MAJOR but 1 as E_TSP_REVISION, naming the version found — Rev 3 (AAC), its pre-merge draft (ABA) and Rev 2 (AAB) included, all of which this package used to read.
• The sender VID is required in every confidential payload (spec §3.2, §3.7 step 7). Rev 3 allowed the NULL VID under HPKE-Base; a NULL sender field is now refused. Packing is unchanged — we always wrote the VID.
• The Rev 2 reader is deleted (the "one rm" CLAUDE.md anticipated), with its dispatcher arm, UnpackKeys.senderEncryptionKey, the revision field/type and the Rev 2 cesr constants. src/rev3/ → src/v1/. KNOWN_MINORS → KNOWN_VERSIONS; describeVersion added. CLAUDE.md's "TSP Rev 2 exception" paragraph now says the exception is gone and why.
• core takes ^0.4.0 and stops passing the Rev 2-only sender X25519 key.

Not backwards compatible on the wire. The wallet and any VTA it talks TSP to must both be on 1.0 — VTA on affinidi-tsp 0.3 / affinidi-messaging-sdk 0.39. A VTA on 0.2.x is refused by name (E_TSP_REVISION), and refuses ours.

Evidence

• Appendix A 1.0 vectors (tests/fixtures/spec-v1-vectors.json, identical to affinidi-tsp 0.3.0's, checked against the spec text and tsp_sdk 0.16.0's JSON): all HPKE-Base vectors open; the six we can pack re-pack byte for byte; new control-rfi-referral opened and its Signature_new verified.
• Both directions against the published affinidi-tsp 0.3.0: a message it packed opens here with a matching thread digest (tests/interop.rust-vector.mjs), and a message packed here with the same fixed keys was opened by its direct::unpack — payload, sender, receiver, digest all match.
• npm run lint, npm run build, npm test all green across workspaces: tsp-js 123/123, core 1070 pass / 0 fail (2 pre-existing schema skips), extension 1308/1308.

Release: @openvtc/vti-tsp-js 0.3.0 → 0.4.0 (needs publishing for core's ^0.4.0 to resolve outside the workspace). @openvtc/pnm-core has an Unreleased CHANGELOG entry.

Pre-merge checklist

• No new reqwest::Client::new() / bare fetch(); all clients have finite timeouts (R1.2) — no network code touched
• No lock held across a network await (R1.3) — N/A
• No local state committed before its remote effect (R2.1) — N/A, codec change
• Every retry is bounded + backed off (R1.4) — N/A
• Accept/poll/listen loops survive transient errors (R1.5) — N/A; a draft-version inbound frame surfaces as a typed E_TSP_REVISION like any other refused frame
• Acks/deletes happen only after durable handoff (R1.6) — N/A
• New/changed wire types (R3.*) — TSP wire 0.2 → 1.0 and a stricter receive rule; matches affinidi-tsp 0.3.0 (already published), verified in both directions
• Config absence = most restrictive (R5.*) — unknown MAJOR and NULL confidential sender both refused
• Logs/status claim only what was verified (R6.*) — refusals carry the version found
• "Process dies on the next line" (R2.1) — no mutations
• Deviations flagged: R3.6 — wire-breaking; the VTA side shipped in affinidi-tdk-rs#967. No compatibility fold kept, per this repo's CLAUDE.md