import InductiveVerification.Public -- The Needham-Schroeder Public-Key Protocol namespace NS_Public variable [InvKey] variable [Bad] open Msg open Event open Bad open HasInitState open InvKey attribute [-simp] Key.injEq -- Define the inductive set `ns_public` inductive ns_public : List Event → Prop | Nil : ns_public [] | Fake : ns_public evsf → X ∈ synth (analz (spies evsf)) → ns_public (Says Agent.Spy B X :: evsf) | NS1 : ns_public evs1 → Nonce NA ∉ used evs1 → ns_public (Says A B (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄) :: evs1) | NS2 : ns_public evs2 → Nonce NB ∉ used evs2 → Says A' B (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄) ∈ evs2 → ns_public (Says B A (Crypt (pubEK A) ⦃Nonce NA, Nonce NB, Agent B⦄) :: evs2) | NS3 : ns_public evs3 → Says A B (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄) ∈ evs3 → Says B' A (Crypt (pubEK A) ⦃Nonce NA, Nonce NB, Agent B⦄) ∈ evs3 → ns_public (Says A B (Crypt (pubEK B) (Nonce NB)) :: evs3) -- A "possibility property": there are traces that reach the end theorem possibility_property : ∃ NB, ∃ evs, ns_public evs ∧ Says A B (Crypt (pubEK B) (Nonce NB)) ∈ evs := by exists 1 exists [ Says A B (Crypt (pubEK B) (Nonce 1)), Says B A (Crypt (pubEK A) ⦃Nonce 0, Nonce 1, Agent B⦄), Says A B (Crypt (pubEK B) ⦃Nonce 0, Agent A⦄), ] constructor · apply ns_public.NS3 · apply ns_public.NS2 · apply_rules[ns_public.NS1, ns_public.Nil, Nonce_notin_used_empty] · simp · tauto all_goals tauto · simp -- Spy never sees another agent's private key unless it's bad at the start @[simp] theorem Spy_see_priEK {h : ns_public evs} : (Key (priEK A) ∈ parts (spies evs)) ↔ A ∈ bad := by constructor · induction h with | Nil => simp[spies, knows, initState, pubEK, priEK, pubSK, Key.injEq] | Fake _ h => apply Fake_parts_sing at h intro h₁; simp at h₁; apply Fake_parts_sing_helper (h := h) at h₁ simp_all | NS1 => simp_all | NS2 => simp_all | NS3 => simp_all · intro _; apply_rules [ parts_increasing, Spy_spies_bad_privateKey ] @[simp] theorem Spy_analz_priEK {h : ns_public evs} : Key (priEK A) ∈ analz (spies evs) ↔ A ∈ bad := by constructor · intro h₁; apply analz_subset_parts at h₁; aapply Spy_see_priEK.mp · intro h₁; apply analz_increasing; aapply Spy_spies_bad_privateKey -- It is impossible to re-use a nonce in both NS1 and NS2, provided the nonce is secret theorem no_nonce_NS1_NS2 { evs: List Event} { h : ns_public evs } : (Crypt (pubEK C) ⦃NA', Nonce NA, Agent D⦄ ∈ parts (spies evs) → (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄ ∈ parts (spies evs) → Nonce NA ∈ analz (spies evs))) := by intro h₁ h₂ induction h with | Nil => simp[spies, knows] at h₂ | Fake _ h ih => apply analz_spies_mono simp [*] at * apply Fake_parts_sing at h apply Fake_parts_sing_helper (h := h) at h₁ apply Fake_parts_sing_helper (h := h) at h₂ simp [*] at *; grind[analz_subset_parts] | NS1 _ nonce_not_used => apply analz_spies_mono simp [*] at * expand_parts_element at h₁; expand_parts_element at h₂; grind [ parts_knows_Spy_subset_used ] | NS2 _ nonce_not_used => apply analz_spies_mono simp [*] at * expand_parts_element at h₂; grind [ parts_knows_Spy_subset_used ] | NS3 _ _ _ a_ih => apply analz_spies_mono; simp_all -- Unicity for NS1: nonce NA identifies agents A and B theorem unique_NA { h : ns_public evs } : (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄ ∈ parts (spies evs) → (Crypt (pubEK B') ⦃Nonce NA, Agent A'⦄ ∈ parts (spies evs) → (Nonce NA ∉ analz (spies evs) → A = A' ∧ B = B'))) := by induction h with | Nil => simp[spies, knows] | Fake _ a a_ih => apply Fake_parts_sing at a; intro h₁ h₂ h₃; apply analz_spies_mono_neg at h₃; simp [*] at * apply Fake_parts_sing_helper (h := a) at h₁ apply Fake_parts_sing_helper (h := a) at h₂ simp_all | NS1 => intro h₁ h₂ _; simp [*] at * expand_parts_element at h₁; expand_parts_element at h₂ grind [ analz_insert_mono_neg, parts_knows_Spy_subset_used_neg ] | NS2 => intro _ _ h₃; apply analz_insert_mono_neg at h₃; simp_all | NS3 => intro _ _ h₃; apply analz_insert_mono_neg at h₃; simp_all; -- Spy does not see the nonce sent in NS1 if A and B are secure theorem Spy_not_see_NA { h : ns_public evs } { not_bad_A : A ∉ bad } { not_bad_B : B ∉ bad } : Says A B (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄) ∈ evs → Nonce NA ∉ analz (spies evs) := by intro h₁ h₄ induction h with | Nil => simp_all | Fake _ a => have _ := Spy_in_bad; apply Fake_analz_insert at a; apply a at h₄; simp_all | NS1 _ a a_ih => simp_all; rcases h₁ with (_ | h) · simp_all; apply a; aapply analz_knows_Spy_subset_used · apply analz_insert_Crypt_subset at h₄; simp at h₄; cases h₄ <;> simp_all apply Says_imp_used at h; apply used_parts_subset_parts at h simp_all[Set.subset_def] | NS2 _ _ a a_ih => simp [*] at *; have _ := h₄; have c := h₁ apply Says_imp_parts_knows_Spy at h₁ have d := h₁ expand_parts_element at d apply analz_insert_Crypt_subset at h₄; simp at h₄; rcases h₄ with (h |h |h) <;> simp [*] at *; · apply a_ih at c; have _ := c; apply Says_imp_parts_knows_Spy at a apply unique_NA at h₁; apply h₁ at a; apply a at c; all_goals simp_all · grind[parts_knows_Spy_subset_used] | NS3 => apply analz_insert_Crypt_subset at h₄; simp[*] at h₄; grind [Says_imp_parts_knows_Spy, no_nonce_NS1_NS2] -- Authentication for `A`: if she receives message 2 and has used `NA` to start a run, then `B` has sent message 2. theorem A_trusts_NS2 {h : ns_public evs } { not_bad_A : A ∉ bad } { not_bad_B : B ∉ bad } : Says A B (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄) ∈ evs → Says B' A (Crypt (pubEK B) ⦃Nonce NA, Nonce NB, Agent B⦄) ∈ evs → Says B A (Crypt (pubEK B) ⦃Nonce NA, Nonce NB, Agent B⦄) ∈ evs := by intro h₁ h₂; apply Says_imp_parts_knows_Spy at h₂ -- use unique_NA to show that B' = B induction h with | Nil => simp_all | Fake _ a => have snsNA := h₁; apply Spy_not_see_NA at snsNA <;> try assumption apply analz_spies_mono_neg at snsNA simp [*] at * cases h₁ · simp_all[Spy_in_bad] · apply Fake_parts_sing at a; apply Fake_parts_sing_helper (h := a) at h₂; simp at h₂ grind [analz_subset_parts] · aapply ns_public.Fake | NS1 => simp [*] at *; expand_parts_element at h₂ grind[parts_knows_Spy_subset_used_neg] | NS2 => simp [*] at * grind [ Spy_not_see_NA, Says_imp_parts_knows_Spy, unique_NA ] | NS3 => simp_all; -- If the encrypted message appears then it originated with Alice in `NS1` lemma B_trusts_NS1 { h : ns_public evs} : Crypt (pubEK B) ⦃Nonce NA, Agent A⦄ ∈ parts (spies evs) → Nonce NA ∉ analz (spies evs) → Says A B (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄) ∈ evs := by intro h₁ h₂ induction h with | Nil => simp[spies, knows] at h₁ | Fake _ a a_ih => apply analz_spies_mono_neg at h₂ simp at h₁; apply Fake_parts_sing at a; apply Fake_parts_sing_helper (h := a) at h₁; simp_all | NS1 => apply analz_spies_mono_neg at h₂; simp_all; cases h₁ <;> simp_all | NS2 => apply analz_spies_mono_neg at h₂; simp_all; | NS3 => apply analz_spies_mono_neg at h₂; simp_all; -- Authenticity Properties obtained from `NS2` -- Unicity for `NS2`: nonce `NB` identifies nonce `NA` and agent `A` theorem unique_NB { h : ns_public evs } : (Crypt (pubEK A) ⦃Nonce NA, Nonce NB, Agent B⦄ ∈ parts (spies evs) → (Crypt (pubEK A') ⦃Nonce NA', Nonce NB, Agent B'⦄ ∈ parts (spies evs) → (Nonce NB ∉ analz (spies evs) → A = A' ∧ NA = NA' ∧ B = B'))) := by -- Proof closely follows that of unique_NA induction h with | Nil => aesop (add norm spies, norm knows, safe analz_insertI) | Fake _ a a_ih => apply Fake_parts_sing at a; intro h₁ h₂ h₃; simp [*] at * apply Fake_parts_sing_helper (h := a) at h₁; apply Fake_parts_sing_helper (h := a) at h₂; simp [*] at * apply analz_insert_mono_neg at h₃ grind[analz_subset_parts] | NS1 => intro _ _ h₃; apply analz_spies_mono_neg at h₃; simp_all | NS2 => intro h₁ h₂ _; simp [*] at * expand_parts_element at h₁ expand_parts_element at h₂ grind[analz_insert_mono_neg, parts_knows_Spy_subset_used] | NS3 => intro _ _ _; simp_all; grind[analz_insert_mono_neg] -- `NB` remains secret theorem Spy_not_see_NB { h : ns_public evs } { not_bad_A : A ∉ bad } { not_bad_B : B ∉ bad } : Says B A (Crypt (pubEK A) ⦃Nonce NA, Nonce NB, Agent B⦄) ∈ evs → Nonce NB ∉ analz (spies evs) := by intro h₁ h₄ induction h with | Nil => simp_all | Fake _ a => apply Fake_analz_insert at a; apply a at h₄; simp_all[Spy_in_bad]; | NS1 => simp [*] at * apply analz_insert_Crypt_subset at h₄; simp at h₄ have h₂ := h₁; apply Says_imp_parts_knows_Spy at h₂ expand_parts_element at h₂ grind[parts_knows_Spy_subset_used] | NS2 => simp [*] at * have _ := h₄ apply analz_insert_Crypt_subset at h₄; simp at h₄ rcases h₁ with (_ | h₁) · simp_all; grind[parts_knows_Spy_subset_used, analz_subset_parts] · have _ := h₁; apply Says_imp_parts_knows_Spy at h₁ expand_parts_element at h₁ grind[ parts_knows_Spy_subset_used, Says_imp_parts_knows_Spy, no_nonce_NS1_NS2 ]; | NS3 => simp at h₁; simp[analz_insert_Crypt_element] at h₄; rcases h₄ with (⟨_, _⟩ | ⟨_, _⟩) <;> simp_all grind[Says_imp_parts_knows_Spy, unique_NB] -- Authentication for `B`: if he receives message 3 and has used `NB` in message -- 2, then `A` has sent message 3. theorem B_trusts_NS3 { h : ns_public evs } { not_bad_A : A ∉ bad } { not_bad_B : B ∉ bad } : Says B A (Crypt (pubEK A) ⦃Nonce NA, Nonce NB, Agent B⦄) ∈ evs → Says A' B (Crypt (pubEK B) (Nonce NB)) ∈ evs → Says A B (Crypt (pubEK B) (Nonce NB)) ∈ evs := by intro h₁ h₂ apply Says_imp_parts_knows_Spy at h₂ induction h with | Nil => simp_all | Fake _ a a_ih => simp [*] at * apply Fake_parts_sing at a apply Fake_parts_sing_helper (h := a) at h₂; simp at h₂ grind [ Spy_in_bad, analz_subset_parts, Spy_not_see_NB ] | NS1 => simp_all | NS2 => simp [*] at *; expand_parts_element at h₂ grind[ parts_knows_Spy_subset_used ]; | NS3 => simp [*] at *; grind [Spy_not_see_NB, Says_imp_parts_knows_Spy, unique_NB] -- Overall guarantee for `B` -- If NS3 has been sent and the nonce NB agrees with the nonce B joined with NA, -- then A initiated the run using NA theorem B_trusts_protocol { h : ns_public evs } { not_bad_A : A ∉ bad } { not_bad_B : B ∉ bad } : Crypt (pubEK B) (Nonce NB) ∈ parts (spies evs) → Says B A (Crypt (pubEK A) ⦃Nonce NA, Nonce NB, Agent B⦄) ∈ evs → Says A B (Crypt (pubEK B) ⦃Nonce NA, Agent A⦄) ∈ evs := by intro h₁ h₂ induction h with | Nil => simp_all | Fake _ a a_ih => simp [*] at * apply Fake_parts_sing at a apply Fake_parts_sing_helper (h := a) at h₁; expand_parts_element at h₁ grind[Spy_in_bad, analz_subset_parts, Spy_not_see_NB] | NS1 => simp_all | NS2 _ nonce_not_used a a_ih => simp [*] at *; expand_parts_element at h₁ grind[parts_knows_Spy_subset_used]; | NS3 => simp [*] at * grind[Spy_not_see_NB, Says_imp_parts_knows_Spy, unique_NB ] end NS_Public