fips205-slhdsa-verified/verification/Proofs/Audit.lean

151 lines
8.7 KiB
Text
Raw Normal View History

review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
/- ──────────────────────────────────────────────────────────────────────────────
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
Proofs/Audit.lean — the axiom-cone + statement + coverage audit, in Lean.
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
History: round 2 moved the axiom check inside Lean (collectAxioms, exact
per-cert set equality). Round 4 (third reviewer, 2026-07-24) demonstrated
that exact-cone-per-cert, while sound FOR LISTED certs, left three gaps
OUTSIDE the cone check — all fail-open:
F1 the audited SET was unbound: deleting a manifest row silently drops a
cert, and adding an un-manifested theorem (e.g. `: False := cheat _`)
was never looked at → the repo could prove False under ALL GREEN.
F2 only cones were bound, never STATEMENTS: a certificate whose type was
replaced by a tautology of the same cone passed green.
F3 (handled in check.sh) the gen/ model bytes were unbound to provenance.
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
This file now closes F1 and F2 with three layers:
(1) NAMED CERTS — each of the eleven: exists ∧ is a `theorem` ∧ cone ==
its expected set EXACTLY ∧ its elaborated-type structural fingerprint
(`Expr.hash`) == the committed value. A gutted statement changes the
fingerprint → fail (F2).
(2) FULL-MODULE ENUMERATION — EVERY theorem defined in the eight
certificate modules must have a cone ⊆ the allowed boundary. An
un-manifested `: False := cheat _` has cone {cheat} ⊄ boundary → fail,
whether or not anyone "listed" it (F1, the dangerous half).
(3) MANIFEST FINGERPRINT — a hash over the whole committed manifest
(names + cones + type fingerprints) is printed; check.sh binds to the
exact committed value, so deleting/swapping a row, or editing a cone or
a fingerprint, changes it and fails the build outside Lean too (F1, the
set-binding half).
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
Any mismatch is a `throwError` → non-zero `lean` exit. Nothing fails open for
a listed cert, an unlisted theorem, a gutted statement, or a dropped oracle.
The standing limit (unchanged, disclosed): an audit cannot defend against an
author who edits the manifest AND check.sh AND the proofs together; the
consumer defense is the pinned commit reviewed at the pin.
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
────────────────────────────────────────────────────────────────────────────── -/
import Proofs.ChainSpec
import Proofs.WotsSpec
import Proofs.XmssSpec
import Proofs.HtSpec
import Proofs.ForsInnerSpec
import Proofs.ForsOuterSpec
import Proofs.InputPrepSpec
import Proofs.ApexSpec
open Lean Elab Command
namespace SlhVerify.Audit
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
/-- Lean's three kernel axioms. -/
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
def kernel3 : List Name := [`propext, `Classical.choice, `Quot.sound]
def oracleF : Name := `verify_mono.oracle.f
def oracleH : Name := `verify_mono.oracle.h
def oracleTL : Name := `verify_mono.oracle.t_l
def oracleTLen : Name := `verify_mono.oracle.t_len
def oracleHMsg : Name := `verify_mono.oracle.h_msg
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
/-- The only axioms permitted in ANY cone in the certificate modules: kernel-3
plus the five SHA-2 oracles. -/
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
def allowedBoundary : List Name :=
kernel3 ++ [oracleF, oracleH, oracleTL, oracleTLen, oracleHMsg]
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
/-- The eight certificate-bearing proof modules. Every `theorem` defined in
these must have a cone ⊆ allowedBoundary (layer 2). -/
def certModules : List Name :=
[`Proofs.ChainSpec, `Proofs.WotsSpec, `Proofs.XmssSpec, `Proofs.HtSpec,
`Proofs.ForsInnerSpec, `Proofs.ForsOuterSpec, `Proofs.InputPrepSpec, `Proofs.ApexSpec]
/-- THE COMMITTED MANIFEST: for each certificate, its exact expected cone and
the structural fingerprint (`Expr.hash`) of its elaborated statement. Ground
truth captured 2026-07-27 via Probe.lean; cross-checked against the round-1/2
cone reconstructions. Changing a proof (cone) OR a statement (fingerprint)
breaks the audit; adding/removing a row changes the manifest fingerprint that
check.sh binds to. -/
def manifest : List (Name × List Name × UInt64) :=
[ (`fips205.chain_free_loop_eq, kernel3 ++ [oracleF], 2535491171),
(`fips205.wots_loop1_eq, kernel3 ++ [oracleF], 2968777228),
(`fips205.xmss_loop_eq, kernel3 ++ [oracleH], 763493610),
(`fips205.ht_loop_eq, kernel3 ++ [oracleF, oracleH, oracleTL], 1071622136),
(`fips205.fors_inner_loop_eq, kernel3 ++ [oracleH], 2102092697),
(`fips205.fors_outer_loop_eq, kernel3 ++ [oracleF, oracleH], 2749583942),
(`fips205.to_int_loop_eq, kernel3, 3797401829),
(`fips205.to_byte_loop_eq, kernel3, 3618049364),
(`fips205.wots_csum_loop_eq, kernel3, 4245103433),
(`fips205.base2b_outer_loop_eq, kernel3, 324621577),
(`fips205.slh_verify_128s_accepts_iff, kernel3 ++ [oracleF, oracleH, oracleTL, oracleTLen, oracleHMsg], 2489587792) ]
/-- Canonical serialization of the committed manifest → one hash. check.sh binds
to the printed value, so any row add/remove/edit, cone change, or fingerprint
change flips it and fails the build. -/
def manifestFingerprint : UInt64 :=
let sortName (l : List Name) : List Name :=
((l.map toString).toArray.qsort (· < ·)).toList.map (·.toName)
let ser := String.intercalate ";" (manifest.map (fun (n, cone, fp) =>
s!"{n}|{String.intercalate "," ((sortName cone).map toString)}|{fp}"))
String.hash ser
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
elab "auditCones" : command => do
let env ← getEnv
let mut errs : Array String := #[]
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
-- (0) the manifest's expected cones may reference nothing outside the boundary
for (cert, cone, _) in manifest do
for a in cone do
unless allowedBoundary.contains a do
throwError "manifest references non-boundary axiom {a} for {cert}"
-- (1) NAMED CERTS: exists ∧ theorem ∧ exact cone ∧ statement fingerprint
for (cert, expected, expFp) in manifest do
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
match env.find? cert with
| none => errs := errs.push s!"{cert}: NOT FOUND (renamed/deleted?)"
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
| some (.thmInfo ci) =>
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
let got := (← collectAxioms cert).toList
let extras := got.filter (fun a => !expected.contains a)
let missing := expected.filter (fun a => !got.contains a)
unless extras.isEmpty && missing.isEmpty do
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
errs := errs.push s!"{cert}: cone extra={extras} missing={missing}"
unless ci.type.hash == expFp do
errs := errs.push s!"{cert}: STATEMENT fingerprint {ci.type.hash} ≠ committed {expFp} (statement changed?)"
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
| some (.axiomInfo _) => errs := errs.push s!"{cert}: is an AXIOM, not a proven theorem"
| some (.opaqueInfo _) => errs := errs.push s!"{cert}: is OPAQUE, not a proven theorem"
| some _ => errs := errs.push s!"{cert}: not a theorem"
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
-- (2) FULL-MODULE ENUMERATION: every theorem in a cert module has a clean cone.
-- This is the layer that stops an un-manifested `: False := cheat _`.
let manifestNames := manifest.map (·.1)
let mut nModuleThms := 0
let mut certsSeen : Array Name := #[]
for (nm, ci) in env.constants.toList do
match ci with
| .thmInfo _ =>
match env.getModuleIdxFor? nm with
| some idx =>
if certModules.contains env.header.moduleNames[idx.toNat]! then
nModuleThms := nModuleThms + 1
if manifestNames.contains nm then certsSeen := certsSeen.push nm
let cone := (← collectAxioms nm).toList
let bad := cone.filter (fun a => !allowedBoundary.contains a)
unless bad.isEmpty do
errs := errs.push s!"UN-AUDITED theorem {nm} (module {env.header.moduleNames[idx.toNat]!}) has disallowed axioms {bad}"
| none => pure ()
| _ => pure ()
-- every manifest cert must actually be a theorem found in a cert module
for cert in manifestNames do
unless certsSeen.contains cert do
errs := errs.push s!"manifest cert {cert} not found as a theorem in any certificate module"
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
unless errs.isEmpty do
review round 4: bind the cert set, statements, and model bytes (F1/F2/F3) The third reviewer demonstrated that the round-2 in-Lean exact-cone audit, though sound for LISTED certs, left three fail-opens OUTSIDE the cone check — and made check.sh print ALL GREEN over a repo proving False. All closed; no theorem, proof, or fold changed (the 11 cones are unchanged). F1 — the audited SET was unbound. Audit.lean now (a) enumerates EVERY theorem defined in the eight certificate modules and requires each cone ⊆ boundary, so an un-manifested `theorem _ : False := cheat _` fails regardless of naming (this is the exact exploit the reviewer used); and (b) prints a MANIFEST fingerprint over the whole committed manifest, which check.sh binds to — so deleting/swapping a cert row fails outside Lean too. F2 — only cones were bound, not statements. Each cert now also carries the structural fingerprint (Expr.hash) of its elaborated type; a statement gutted to a tautology of the same cone changes the fingerprint and fails. F3 — the gen/ model bytes were unbound. New check.sh Phase 0 sha256-pins all four gen/SlhVerify/*.lean (incl. the two hand-maintained *External files, now hashed in PROVENANCE.json) BEFORE compiling; a hand-edited model fails first. F4/F5 — docs. README cone diagram now roots honestly at slh_verify_internal and states the pure/prehash domain-separator byte, the ctx>255 check, M' assembly, and deserialization are ABOVE the root and uncovered (new TRUSTED-BASE item 10). The false "rules out a wrong ADRS field" claim is corrected in README + ChainSpec (a transliteration makes the field visible, not excluded). check-selftest.sh: eight attacks, all rejected (dead file; extra axiom; dropped oracle; vanished cert; un-manifested False theorem; gutted statement; hand-edited model; deleted manifest row). Full transcript + green check.sh in verification/RECORDED-RUN.md. Standing limit unchanged and disclosed: an audit cannot defend against an author who edits the manifest AND check.sh AND the proofs together; the consumer defense is the pinned commit reviewed at the pin. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-27 17:47:39 +00:00
throwError "AUDIT FAILED (fail-closed):\n{String.intercalate "\n" errs.toList}"
logInfo s!"exact-cone audit PASSED: {manifest.length} certificates (cones + statement fingerprints), {nModuleThms} module theorems enumerated clean; MANIFEST-FINGERPRINT: {manifestFingerprint}"
review round 2: in-Lean exact-cone audit + reproducibility + doc honesty Addresses the round-2 reviewer punch-list. No theorem statement, proof term, or fold definition changed; the eleven cones are unchanged (independent collectAxioms dump in verification/RECORDED-RUN.md). AUDIT GATE (both reviewers, the critical one) - Retire the bash #print-axioms text parser (fail-open on empty/truncated reports, and only a SUBSET check). Replace with verification/Proofs/Audit.lean: reads each certificate's cone from the kernel via collectAxioms and asserts EXACT set equality against its expected boundary. Extra axiom, dropped oracle, renamed/deleted cert, or an axiom/opaque sham each throw -> non-zero Lean exit. No text to misparse; nothing fails open. check.sh Phase 3 now just compiles it (and still requires the explicit PASSED line). - check-selftest.sh rewritten to attack the new gate: dead-file, smuggled extra axiom (named), dropped-oracle (subset would pass, exact must not), and a vanished certificate (the collectAxioms-returns-[] trap). All four rejected. REPRODUCIBILITY (GPT B1.4 / B1.5) - extract.sh refuses a wrong-commit or dirty source tree (fail-closed), takes an optional source-path arg, and pins the source commit. - verification/PROVENANCE.json: single machine-readable pin set (source + charon + aeneas commits/channel + lean + ocaml) with generated-file sha256. - Re-running extract.sh reproduces gen/SlhVerify/{Types,Funs}.lean byte-identically (companion fips205-source commit adds Cargo.lock + rust-toolchain.toml; verified not to perturb the model). DOC HONESTY (both reviewers) - README: fix the self-contradiction (apex "not yet proven" trailer vs the proven apex), the false "oracles kept OUTSIDE every cone" (they are INSIDE, by design), "deployed monomorphic path" and "semantics-identical for every parameter set" overclaims, "only two lines changed", stale snapshot head; retitle the stale future-tense "what will be claimed" section. - TRUSTED-BASE: drop "nothing proven yet"; add base_2b-inner and deployment- bridge non-claims explicitly; current pin. - ChainSpec header: "deployed monomorphic path" -> private verify_mono facade (comment only). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-24 17:13:55 +00:00
end SlhVerify.Audit
open SlhVerify.Audit in
auditCones