Commit graph

2 commits

Author SHA1 Message Date
3d7c125789 entries 13-18: re-attest the estate at 44 certs/fork + first SLH-DSA leaf; heads now dual-signed
Six new leaves (tree 13 -> 19, root 7ee23940…):

  [13-16] dalek/anza/risc0/betrusted-ed25519-verified — re-attested at 44
          certificates each (27 main + 4 apex + 13 scalar; the leaf 8-11
          generation recorded 16). The delta is the P0-P2 hardening
          campaign: scalar statements bound, kernel-side axiom gate,
          driver surface with cones, declaration coverage both directions,
          the accounting identity.
  [17]    ltl-accumulator-verified — 61 certificates; the log again carries
          proofs of its own Merkle machinery, at the hardened state.
  [18]    fips205-slhdsa-verified — FIRST post-quantum leaf: 11 certificates
          over the SLH-DSA-SHA2-128s verify path, apex
          fips205.slh_verify_128s_accepts_iff.

Heads from tree 14 on carry a second, ADDITIVE signature:
SLH-DSA-SHA2-128s (FIPS 205), deterministic, over the same payload as the
Ed25519 signature. provider.slhdsa.pub ships beside the Ed25519 key.
Ed25519 remains the signature consumers must check; verify.py now judges
slh_dsa fail-closed where present (INVALID/WRONG-KEY/NO-PUBKEY fail; older
heads report ABSENT, allowed; pre-3.5 OpenSSL degrades loudly, never
silently). The ml_dsa slot stays not_configured — truthfully.

Honesty, unchanged by any of it: the certificates cover VERIFICATION paths
of the extracted Lean models; no signing operation is proven for any
algorithm; leaves are Ed25519-signed at issuance only.

Append-only law checked byte-for-byte before this commit: entries
000000-000012 identical, the six prior heads an exact prefix of the
history. verify.py --all: RESULT OK [full]. verify_selftest.py: GREEN (13
cases). Quorum gate: 5-way Ed25519 + 2-way SLH-DSA (incl. the verifier
built from the pinned proven source), all accept, all reject corruption.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-08-07 17:11:59 +02:00
52179bdaf1 verify.py: --all verifies every published receipt; binding fields required, not compare-if-present (round-12 GPT B4)
Round-12 review found two artifact-claim gaps: --all never touched
receipts, and receipt mode checked fingerprint/leaf_hash only when
present. Now:
- --all enumerates receipts/*.receipt.json and runs the full binding
  list on each (type tag, STH signature, REQUIRED key fingerprint,
  log_id vs log-metadata, STH membership in the published history,
  REQUIRED leaf_hash vs the named entry, tree_size agreement, receipt
  hash_algorithm + log_id consistency, inclusion proof).
- --receipt FILE routes through the same function.
- Malformed values (bad hex, missing sizes) are failures, not crashes.
- NEW verify_selftest.py: 11-case adversarial battery (mutated real
  receipts must be REJECTED: no fingerprint, no leaf_hash, wrong type,
  forged root, size mismatch, wrong log_id; honest controls pass;
  structural-only never claims full; no-openssl exits 2). GREEN.
One bug caught by the honest controls during development: the new
hash-algorithm check assumed 'sha256' but deployed heads carry
'RFC9162_SHA256' - fixed against reality, plus receipt-level
hash_algorithm/log_id binding added.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-17 11:15:19 +02:00