review round 1: fix the fail-open audit gate + remove the overclaimed framing

External review (both standing reviewers, 2026-07-24) returned DO NOT ATTEST.
The eleven Lean theorems compile with genuinely clean cones (both reviewers
independently reconstructed them), but two real defects were found and are
fixed here.

FIX 1 — the axiom audit was FAIL-OPEN (the critical blocker). check.sh Phase 3
grepped a single physical line of each `#print axioms` report; Lean WRAPS long
cones across lines, so for ht/fors_outer/APEX the audit checked only `[propext,`
and silently ignored the continuation lines — a disallowed axiom on line 2+
passed (the GPT reviewer demonstrated `review_evil_ax` passing). Since check.sh
is the sole source of the word "proven", this is unacceptable.
  - New parser: FLATTEN the whole report (join newlines) BEFORE parsing, then
    extract each certificate's complete bracketed cone with a literal-string
    (regex-safe) scan and subset-check every axiom. Missing/empty report => FAIL
    CLOSED. The audit now prints the count of axioms actually audited per cert
    (apex: 8, previously 1).
  - check-selftest.sh gains ATTACK 3: a smuggled axiom bundled with the apex so
    its cone WRAPS with the evil axiom on a continuation line — the exact
    exploit. Verified: all three attacks now rejected, attack 3 via the axiom
    gate naming the continuation-line axiom. (Also fixed attack 2's leftover
    EvilSpec.lean tripping attack 3's dead-file gate.)

FIX 2 — remove the overclaimed framing (refuted by both reviewers). Corrected
in README, the ApexSpec header + apex docstring, and (separately) the control
MANIFEST:
  - "composes all ten loop-fidelity certificates" — FALSE. The apex proof is a
    STRUCTURAL FACTORIZATION; it references NONE of the ten (grep: 0) and would
    remain provable if one were deleted. They are independent local-fidelity
    lemmas, not links in the apex proof.
  - "every loop is individually fidelity-certified" — FALSE. base_2b's inner
    accumulation loop is threaded opaquely and uncertified — and it determines
    the FORS indices / WOTS digits, so a defect there could change the recomputed
    root while all eleven theorems still hold.
  - "the deployed verifier" — the proved subject is verify_mono, a private
    #![allow(dead_code)] monomorphic facade NOT called by the public API; the
    bridge to the deployed generic verifier is the finite differential test,
    not a machine-checked refinement.
  - "verify-path pyramid complete" — replaced with "intermediate verification
    layer"; the apex is an ACCEPTANCE CHARACTERIZATION, not closed-form FIPS-205
    correctness.
Also: FunsExternal header noted the Take axiom "remains" (stale — deleted in
de-plumbing round 2); corrected.

check.sh green over all eleven certificates under the fixed fail-closed parser
(exit 0, 8 axioms audited for the apex). Nothing about the theorems changed —
they were and are sound; only the audit tool and the claims about them are fixed.

NOT DONE (remaining reviewer blockers, tracked): reproducible extract tuple
(pin commits, de-hard-code extract.sh) + Cargo.lock / toolchain pin. Attestation
remains gated behind review round 2 + the operator halt + the appeal.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

README.md

@@ -5,22 +5,53 @@ path**, extracted from a pure-Rust implementation into Lean 4 via
 Charon/Aeneas — the same pipeline, discipline, and honesty rules as the
 four ed25519 campaigns (`dalek/anza/risc0/betrusted-ed25519-verified`).
 
-## STATUS: THE APEX IS PROVEN — the verify-path pyramid is complete (11 certificates)
+## STATUS: eleven certificates over the extracted verify model (external review round 1 applied)
 
 `verification/check.sh` is **green** (exit 0): the model compiles, the
-proofs compile, and the axiom audit passes. **Eleven theorems across seven
+proofs compile, and the axiom audit passes (the Phase-3 parser was rewritten
-layers — chain (5), WOTS+ (8), XMSS (10), hypertree (12), FORS (17),
+to be fail-closed and wrap-safe after external review round 1, 2026-07-24).
-input-prep (2/3/4), and the apex (20): `fips205.slh_verify_128s_accepts_iff`
+
-— the deployed SLH-DSA-SHA2-128s verifier returns `ok true` if and only if
+**What is actually established** — eleven Lean theorems about the
-the recomputed hypertree root byte-equals the pinned public-key root. Its
+Aeneas-generated model of the **monomorphic `verify_mono` compatibility
-`#print axioms` cone is exactly `[propext, Classical.choice, Quot.sound]`
+verify path** (an additive, `#![allow(dead_code)]` re-expression of the
-plus the five SHA-2 oracles, nothing else. Honest scope: the apex is an
+deployed generic verifier, using named hash oracles because Charon/Aeneas
-acceptance characterization (accept = root equality over the extracted
+cannot translate the deployed `Hashers` function-pointer struct):
-recomputation, whose every loop is individually fidelity-certified by the
+
-ten preceding theorems); it does not restate the recomputation as a
+- **Ten loop-fidelity theorems** (chain 5, WOTS+ 8, XMSS 10, hypertree 12,
-closed-form mathematical hypertree value. After de-plumbing rounds 1+2 the
+  FORS-inner/outer 17, and the input-prep helpers to_int/to_byte/checksum/
-model carries no plumbing axioms on the verify path — its external surface
+  base_2b-outer, Alg 2/3/4). Each equates one *generated* Aeneas loop with an
-is exactly the five SHA-2 oracles (plus off-path zeroize impls).**
+  explicit hand-written recursive fold — a local control-flow correspondence,
+  not an Algorithm-level mathematical specification.
+- **The apex, `fips205.slh_verify_128s_accepts_iff`** — the extracted
+  `verify_mono::slh_verify_128s` returns `ok true` **iff** the recomputed
+  hypertree root byte-equals `pk.pk_root`. This is an *acceptance
+  characterization*: there is no acceptance path other than root equality
+  over the extracted recomputation. Its `#print axioms` cone is exactly
+  `[propext, Classical.choice, Quot.sound]` + the five SHA-2 oracles.
+
+**What is NOT (yet) established — do not overclaim:**
+- **The apex proof does not compose the ten loop theorems.** It is a
+  *structural factorization* of the extracted verifier around its final
+  equality check; it references none of the ten (it would remain provable if
+  one were deleted). The ten are independent local-fidelity lemmas, not links
+  in the apex's proof chain.
+- **Not "every loop":** `base_2b`'s inner accumulation loop
+  (`helpers.base_2b_loop0_loop0`) is threaded *opaquely* and has no
+  certificate — and it determines the FORS indices / WOTS digits, so a defect
+  there could change the recomputed root while all eleven theorems still hold.
+- **Not the deployed public verifier:** the proved subject is the private
+  `verify_mono` facade; the bridge to upstream's generic `pk.verify()` is the
+  finite in-snapshot **differential test**, not a machine-checked refinement.
+- **Not closed-form FIPS 205 correctness:** the folds are transliterations of
+  the extracted loops (the hash primitives stay opaque); nothing here relates
+  the recomputed root to a mathematical SLH-DSA specification.
+
+This is a real **intermediate** verification layer, not an end-to-end
+formal verification of the deployed verifier. After de-plumbing rounds 1+2
+the model carries no plumbing axioms on the verify path — its external
+surface is exactly the five SHA-2 oracles (plus off-path zeroize impls).
+The trust base and residual assumptions are stated in
+[TRUSTED-BASE.md](TRUSTED-BASE.md).
 
 - **`fips205.chain_free_loop_eq`** (Algorithm 5, WOTS+ chaining): the
   extracted `chain_free` loop equals the explicit s-fold hash chain, with

verification/Proofs/ApexSpec.lean

@@ -1,19 +1,34 @@
 /- Proofs/ApexSpec.lean — the APEX certificate.
 
-   THEOREM slh_verify_128s_accepts_iff: the extracted top-level verifier
+   THEOREM slh_verify_128s_accepts_iff: the extracted verifier
-   slh_verify_128s returns `ok true` if and only if the recomputed hypertree
+   `verify_mono::slh_verify_128s` (a private, additive, `#![allow(dead_code)]`
-   root byte-equals the pinned public-key root pk.pk_root. Everything the
+   monomorphic re-expression of the deployed generic verify path — NOT the
-   verifier does after recomputing the root is exactly that byte comparison —
+   public `pk.verify()`, which it is not called by) returns `ok true` if and
-   there is no other acceptance path. The recomputation `slhVerifyRoot` is the
+   only if the recomputed hypertree root byte-equals the pinned public-key root
-   extracted pipeline (H_msg digest → md/idx_tree/idx_leaf split via to_int and
+   pk.pk_root. Everything the verifier does after recomputing the root is
-   masks → fors_pk_from_sig → ht recompute over xmss over wots over chain),
+   exactly that byte comparison — there is no other acceptance path. The
-   whose every loop is individually fidelity-certified by the ten preceding
+   recomputation `slhVerifyRoot` is the extracted pipeline (H_msg digest →
-   theorems. #print axioms = kernel + the five SHA-2 oracles, nothing else.
+   md/idx_tree/idx_leaf split → fors_pk_from_sig → ht recompute over xmss over
+   wots over chain). #print axioms = kernel-3 + the five SHA-2 oracles.
+
+   SCOPE — what this does NOT establish (external review round 1, 2026-07-24):
+   • This proof is a STRUCTURAL FACTORIZATION, not a composition. It does NOT
+     invoke any of the ten loop-fidelity theorems (chain_free_loop_eq, …,
+     base2b_outer_loop_eq); it would remain provable if one were deleted. Those
+     ten are independent local-fidelity lemmas, not links in this proof chain.
+   • NOT "every loop": `base_2b`'s inner accumulation loop is threaded opaquely
+     and has no certificate — yet it determines the FORS indices / WOTS digits,
+     so a defect there could change the recomputed root while this theorem holds.
+   • NOT the deployed public verifier: the bridge from `verify_mono` to the
+     generic `pk.verify()` is the finite in-snapshot differential test, not a
+     machine-checked refinement.
+   • NOT closed-form FIPS 205 correctness: the folds are transliterations of the
+     extracted loops with the hash primitives opaque.
 
    The one real lemma is arrayEqU8_spec: the library array equality
    `PartialEqArray.eq PartialEqU8` on two Array U8 N returns exactly the
    decidable byte-equality of their underlying lists (a List.allM induction).
-   Everything else is unfolding the straight-line composition and threading the
+   Everything else is unfolding the extracted verifier and threading the
    recomputation identically on both sides with bind_congr.
 -/
 import Proofs.InputPrepSpec
@@ -184,9 +199,12 @@ theorem slh_verify_internal_accepts_iff {A D HP K LEN N : Std.Usize} (H M : Std.
   -- (no whnf of the nested ht_verify_free_loop — the ForsOuter lesson).
   simp only [ht_verify_free_split, bind_assoc]
 
-/-- **APEX (deployed SHA2-128s entry).** slh_verify_128s accepts iff the
+/-- **APEX (SHA2-128s facade entry).** The extracted `verify_mono::slh_verify_128s`
-    recomputed root byte-equals pk.pk_root. Composes all ten loop-fidelity
+    accepts iff the recomputed root byte-equals pk.pk_root. Acceptance
-    certificates through the extracted pipeline. -/
+    characterization only — see the file header for the four explicit
+    non-claims (this is a structural factorization, NOT a composition of the
+    ten loop certs; NOT the deployed public verifier; base_2b inner uncertified;
+    NOT closed-form FIPS-205 correctness). -/
 theorem slh_verify_128s_accepts_iff
     (mprime : Slice Std.U8)
     (sig : types.SlhDsaSig 12#usize 7#usize 9#usize 14#usize 35#usize 16#usize)

verification/check-selftest.sh

@@ -1,18 +1,24 @@
 #!/usr/bin/env bash
 # Adversarial self-test of the check.sh gates (the R3-5 tradition: an audit
-# that cannot fail is theater). Two attacks, both MUST make check.sh fail:
+# that cannot fail is theater). Three attacks, all MUST make check.sh fail:
 #
 #   1. DEAD FILE   — a stray Proofs/*.lean not in the manifest.
-#   2. SMUGGLED AXIOM — a certificate whose cone contains an axiom outside
+#   2. SMUGGLED AXIOM (short cone) — an axiom outside {kernel-3} ∪ {5 oracles},
-#      {kernel-3} ∪ {the five SHA-2 oracles}.
+#      on the FIRST line of the cone.
+#   3. SMUGGLED AXIOM (WRAPPED cone) — an axiom on a CONTINUATION line of a
+#      cone that wraps (the exact fail-open exploit external review found on
+#      2026-07-24: the old single-line parser saw only line 1). This attack
+#      guards the flattened-parse fix; a self-test that only plants short cones
+#      cannot detect a wrapped-cone parser regression.
 #
-# Green here means: the gates genuinely reject both. Self-cleaning.
+# Green here means: the gates genuinely reject all three. Self-cleaning.
 set -euo pipefail
 HERE="$(cd "$(dirname "$0")" && pwd)"
 cd "$HERE"
 
 cleanup() { rm -f Proofs/Stray.lean Proofs/Stray.olean Proofs/EvilSpec.lean \
-                  Proofs/EvilSpec.olean check-evil-tmp.sh; }
+                  Proofs/EvilSpec.olean Proofs/EvilWrapSpec.lean \
+                  Proofs/EvilWrapSpec.olean check-evil-tmp.sh check-evilwrap-tmp.sh; }
 trap cleanup EXIT
 
 echo "check-selftest: attacking the gates"
@@ -54,7 +60,46 @@ if ./check-evil-tmp.sh > /tmp/selftest-evil.out 2>&1; then
 fi
 grep -q "DISALLOWED" /tmp/selftest-evil.out \
   || { echo "✗ ATTACK 2: failed, but not via the axiom gate"; exit 1; }
+rm -f Proofs/EvilSpec.lean Proofs/EvilSpec.olean check-evil-tmp.sh
 echo "✓ attack 2 rejected (axiom gate works)"
 
+# ── Attack 3: smuggled axiom on a WRAPPED cone (the fail-open exploit) ───────
+# evil_wrapped_thm bundles a disallowed axiom with the apex theorem, so its cone
+# is 9 axioms and WRAPS across physical lines with review_evil_ax on a
+# continuation line — exactly what the old single-line parser missed.
+cat > Proofs/EvilWrapSpec.lean <<'EOF'
+import Proofs.ApexSpec
+open Aeneas Aeneas.Std Result
+open fips205
+axiom review_evil_ax : True
+theorem evil_wrapped_thm
+    (mprime : Slice Std.U8)
+    (sig : types.SlhDsaSig 12#usize 7#usize 9#usize 14#usize 35#usize 16#usize)
+    (pk : types.SlhPublicKey 16#usize) :
+    True ∧ (verify_mono.slh_verify_128s mprime sig pk
+             = (do let root ← slhVerifyRoot 63#usize 30#usize mprime sig pk
+                   ok (decide (root.val = pk.pk_root.val)))) :=
+  ⟨review_evil_ax, slh_verify_128s_accepts_iff mprime sig pk⟩
+EOF
+python3 - <<'PY'
+s = open("check.sh").read()
+assert 'PROOFS=(\n' in s and 'CERTS=(\n' in s, "check.sh array shape changed"
+s = s.replace('PROOFS=(\n', 'PROOFS=(\n  "EvilWrapSpec"\n', 1)
+s = s.replace('CERTS=(\n', 'CERTS=(\n  "evil_wrapped_thm"\n', 1)
+assert '{ echo "import Proofs.ChainSpec"' in s, "check.sh audit import shape changed"
+s = s.replace('{ echo "import Proofs.ChainSpec"',
+              '{ echo "import Proofs.EvilWrapSpec"; echo "import Proofs.ChainSpec"', 1)
+open("check-evilwrap-tmp.sh","w").write(s)
+PY
+chmod +x check-evilwrap-tmp.sh
+if ./check-evilwrap-tmp.sh > /tmp/selftest-evilwrap.out 2>&1; then
+  echo "✗ ATTACK 3 SUCCEEDED: audit passed a smuggled axiom on a WRAPPED cone (fail-open!)"; exit 1
+fi
+grep -q "DISALLOWED" /tmp/selftest-evilwrap.out \
+  || { echo "✗ ATTACK 3: failed, but not via the axiom gate (wrapped-cone parse?)"; exit 1; }
+grep -q "review_evil_ax" /tmp/selftest-evilwrap.out \
+  || { echo "✗ ATTACK 3: rejected, but the audit did not name the continuation-line axiom"; exit 1; }
+echo "✓ attack 3 rejected (wrapped-cone axiom gate works — the continuation-line axiom was seen)"
+
 echo
-echo "SELFTEST GREEN: both gates genuinely reject their attacks."
+echo "SELFTEST GREEN: all three gates genuinely reject their attacks."

verification/check.sh

@@ -90,20 +90,29 @@ AUD="$HERE/Proofs/.audit.lean"
 } > "$AUD"
 OUT=$(lake env bash -c "cd '$HERE' && export LEAN_PATH=\"\$LEAN_PATH:\$PWD/gen:\$PWD\" && LEAN_TIMEOUT=$TIMEOUT LEAN_MEM_MB=$MEM '$HERE/lean-guard' 'Proofs/.audit.lean'" 2>&1)
 rm -f "$AUD"
+# CRITICAL: Lean WRAPS long axiom cones across physical lines. A per-line parse
+# silently drops continuation-line axioms (fail-OPEN — external review round 1,
+# 2026-07-24, demonstrated a smuggled axiom on line 2 passing). FLATTEN the whole
+# report first, then extract each cert's complete bracketed cone with a
+# literal-string (regex-safe) scan and subset-check it. Missing/empty report =>
+# fail-CLOSED.
+FLAT=$(echo "$OUT" | tr '\n' ' ' | tr -s ' ')
 fail=0
 for c in "${CERTS[@]}"; do
-  line=$(echo "$OUT" | grep -F "'$c' depends on axioms:" || true)
+  marker="'$c' depends on axioms: ["
-  if [ -z "$line" ]; then echo "  ✗ $c — no axiom report"; fail=1; continue; fi
+  if ! grep -qF "$marker" <<<"$FLAT"; then
-  cone=$(echo "$line" | sed "s/.*depends on axioms: //")
+    echo "  ✗ $c — no axiom report (fail-closed)"; fail=1; continue
-  # every axiom in the cone must be in ALLOWED
+  fi
-  bad=$(echo "$cone" | tr -d '[]' | tr ',' '\n' | sed 's/^ *//;s/ *$//' | while read -r ax; do
+  # full cone between this cert's '[' and the next ']' (literal index, not regex)
+  cone=$(awk -v m="$marker" '{ i=index($0,m); if(i>0){ r=substr($0,i+length(m)); j=index(r,"]"); if(j>0) print substr(r,1,j-1) } }' <<<"$FLAT")
+  bad=$(echo "$cone" | tr ',' '\n' | sed 's/^ *//;s/ *$//' | while read -r ax; do
     [ -z "$ax" ] && continue
     case " propext Classical.choice Quot.sound verify_mono.oracle.f verify_mono.oracle.h verify_mono.oracle.t_l verify_mono.oracle.t_len verify_mono.oracle.h_msg " in
       *" $ax "*) ;; *) echo "$ax" ;;
     esac
   done)
-  if [ -n "$bad" ]; then echo "  ✗ $c — DISALLOWED axioms: $bad"; fail=1
+  if [ -n "$bad" ]; then echo "  ✗ $c — DISALLOWED axioms: $(echo $bad | tr '\n' ' ')"; fail=1
-  else echo "  ✓ $c  cone ⊆ allowed"; fi
+  else echo "  ✓ $c  cone ⊆ allowed ($(echo "$cone" | tr ',' '\n' | grep -c .) axioms audited)"; fi
 done
 [ "$fail" = 0 ] || { echo "AXIOM AUDIT FAILED"; exit 1; }
 

verification/gen/SlhVerify/FunsExternal.lean

@@ -23,11 +23,14 @@
        Debug-fmt axioms were ELIMINATED at source level by the
        fips205-source de-plumbing patch (8 sites, semantics identical,
        differential-test-validated) and their declarations deleted here
-       (dead-stub rule, 2026-07-23). Remaining as axioms: the Take
+       (dead-stub rule, 2026-07-23). De-plumbing round 2 (2026-07-24) then
-       iterator machinery used by helpers::to_int (slh_verify_internal's
+       removed the LAST iterator adapters: helpers::to_int's Take and base_2b's
-       digest split — the apex round's de-plumbing item) and the zeroize
+       IterMut became index loops, so the Take::next axiom was deleted too.
-       blanket impls (never on the verify path). The #print axioms audit
+       Remaining as axioms on the whole model: the FIVE SHA-2 verify-path
-       confirms only class (1) survives in any certificate cone.
+       oracles and three zeroize blanket impls (never on the verify path) —
+       nothing else. The #print axioms audit (fail-closed, wrap-safe since the
+       external-review fix of 2026-07-24) confirms only the five oracles + the
+       kernel three survive in any certificate cone.
    ────────────────────────────────────────────────────────────────────────────── -/
 -- This is a template file: rename it to "FunsExternal.lean" and fill the holes.
 import Aeneas