Security Threats & Mitigations for Persistent State Issued Citizen Entitlements (SEDI) trade-space walk through Samuel M. Smith, Ph.D. sam@kerisuite.org https://gv'thub.(ozrzglr:loi,t(:SamuelM/Papers
Unique Properties Persistence: potentially life-long timespans over which entitlement needs to verifiable Millions of issuances: Makes mass revocation and re-issuance problematic High value entitlements: Economic assistance, health care, etc. Potential to become the citizen identity assurance root-of-trust for intra- state, inter-state, federal, and private ecosystems Potential to replace existing decentralized identity assurance mechanisms with single universal State endorsement. Persistent, widespread, high value, entitlement system: Highest ROI to attackers = Highest susceptibility to attack. Highest susceptibility means all vulnerabilities will be exploited eventually.
Survivable Systems: AUVs, Ships, Critical Infrastructure —— TET s .? = e “» &"" = > . " 1."'7 ! - v 3 = e = = RIS S | 5 s ) jo? MW‘%% 'M‘f* 7, ey L= s = !'*! T2 ‘ L ’lf) et . .fl.' y fifi
Survivable Security Survivability: Susceptibility - Likelihood of being attacked I Vulnerability - Likelihood of success and extent of harm given an attack ‘x Recoverability - Likelihood and extent of recovery from a successful attack - A system may be highly vulnerable but not highly susceptible to attack s & In IT security, susceptibility stems from the return on investment (ROI) of an attack ¥ The “trap” of confusing low susceptibility with low vulnerability ‘Susoeptbilty e s SRR A Y TN Bl 0\ A St N ey - PR (T Pl Ty .- o " el P B4 = A e - i T L e, P _— P > ,:& = & D % Layered Sunivabiy Overlay Spiral / p
Design Spiral to Explore a Trade Space i 2 \n‘mfd i \ . , iz B weam - e | @i Naval Architecture Design Spiral In harsh trade spaces, incorrect a priori decisions often result in highly suboptimal designs. One must first thoroughly explore the harsh tradeoffs. The PAC trilemma forms a harsh trade space.
PAC Theorem — Trilemma An interaction may be two of the three, private, authentic, and confidential at the highest degree, but not all three at the highest degree. Authentic / Tradespace A / \ & o/ - % L@ A Trade-offs required! s © &
Closed Loop Models for Entitlement Verification Issuer-Holder-Verifier Model with Verification at Issuer Issuance Presentation Issuer Holder Verifier Verification
Open Loop Models for Entitlement Verification Issuer-Holder-Verifier Model with Verification using Verifiable Data Registry Issuance Presentation Issuer Holder Verifier Registration Registration Verifiable Verification Data Registry
Conventional Enterprlse Slgmng Infrastructure Entiitiement Signing Infrastructure | EntilementtobeSigned ~——— Entilementto be Signed Secure Hardware - Access |+ AuthZ Tok ) TPMHSM Control uthz Token Signing L Key Creation and Singing ) Signed Entitiement i ) Signed Entitiement \Requesler' Best practice is to use secure signing hardware (TPM, HSM, Secure Enclave). Asymmetric signing key-pairs created/stored in secure hardware. Private key never leaves hardware. Public key sent elsewhere to be used to verify signature. Role based access control grants AuthZ permissions/tokens to employees who submit entitlements to secure hardware for signing. Weakest vulnerability is the AuthZ infrastructure not the signing hardware. (typically OAuth/OIDC/ 550 based) Systems that do not use secure signing hardware are also vulnerable to direct private key compromise
Attack on Signing Authorization Compromise of the authorization infrastructure results in fraudulent but perfectly verifiable entitlements Typically permissions are repudiable thereby creating plausible deniability Plausible deniability = incentive to suborn employes
Identity Based Attack Statistics 2024 Crowdstrike 2025 report: Breakout time — average fell to 48 minutes Voice phishing — up 442% between the first and second half of 2024. Initial access attacks — 52% of vulnerabilities in 2024. Access broker offerings increased 50% year-over-year. China-nexus activity surged 150%; targeted industries 200% to 300% GenAl enabled infiltration via deep fake IT job candidate and recruiter exploits GenAl Exploit Chaining
Time to Detect and Contain 2024 IBM 2025 data breach report: Average Detection and Containment: In 2024, the total average lifecycle from initial breach to containment wads 194 days to identify and 64 days to contain (258 days). Stolen Credentials: Breaches that begin with stolen or compromised credentials take the longest, with an average lifecycle of 292 days.
Edge Attack on Authorization System OAuth/OIDC, PKI/CA, DNS Hijack, VPN, Remote CLI apps Phishing employee login credentials Al enabled phishing (10x over 2024, 100x over 2023) Recursive privilege escalation attack Stolen bearer tokens (Salesloft Drift->Salesforce->AWS->Google-> ZScaler->PaloAlto) Key compromise impersonation attack Zero Day (share point etc)
Insider Exploit suborn: Disgruntled employee: bribe, revenge, Compromised employee: black mail, familial threats Fake Employee Infiltration: GenAl remote live fakes
Systemic Design Flaw — Risk Issuer Signing Infrastructure is a single point of failure for the whole ecosystem = catastrophic failure potential = annihilation of trust Compromise of that infrastructure could result in rampant impersonation fraud. Exponential increase in adoption amplifies the likelihood of this catastrophic failure Security over issuer signing Infrastructure must therefore employ exponentially stronger security measures
Security First, Always [NOIMORE] HALFIMEASURES] Minimally Sufficient Means Adoptable White Magic Crypto
Inconsistency and Duplicity nconsistency: lacking agreement, s two or more things i relation to each other oy SRS wen dluplicity: acting in two different ways to different people concerning the same matter it | Internal vs. External Inconsistency Rotaton Event . ) " neracton Event Internally inconsistent log = not verifiable. Interacton Event Log verification from a self-certifying cryptographic root-of- Rolation Evert | trust protects against internal inconsistency. Rotaton Event IlractonEvnt | Externally inconsistent logs. Two different logs for the same ol Event —_— identifier, both verifiable = duplicity. Interaction Evert Duplicity detection protects against external inconsistency. KERI provides duplicity evident DKMI
Cate promises to provide a Duplicity Game How may Cate be duplicitous consistent pair-wise log. and not get caught? Local Consistency Guarantee 0 o Different pair with same ID but v oo different log # duplicity / B A LogV1 LogV2 LogV2 a = AP 3 it vaidtor Log Log private (one-to-one) interactions
Service promises to providea Duplicity Game may Cate/Service/Agent be consistent log to anyone. duplicitous and not get caught? Local Consistency Guarantee o Sate & Log I Different pair with same ID but Truncate Log Service/Agent different log # duplicity Contotiet by Cete Delete Log Log LogV1 LogV2 LogV2 Az = 4 Vascator Vardator Log Log highly available, private (one-to-one) interactions
Service promises to provide DUP|ICIty Game ., may Cate andjor service be exact same log to everyone. duplicitous and not get caught? Global Consistency Guarantee 0 Gate Cortrller Global consistency may only matter after Breaking the promise of global VY Log Eve and Joe need to Interact not before. consistency is a provable labilty. I — I Different pair with same ID but Service different log = duplicity Conmoti by Cate .8 | Truncate Log ™ J ? A v Delete Lo A . Doug | Gl upiciy |~ logvi - Lug” — [})g V2 &, | " Ambient Duplicity Detection T e Log - g Log global consistent, highly available, and public (one-to-any) interactions
Recovery from Compromised Signing Infrastructure Entiilement Signing Infrastructure e Key Creation and Singing “Signed Enttement©1 API ‘Sgned Enement Requester Depends on Architecture! Non-specific (prophylactic): Rotate tokens/permissions/keys frequently Closed Loop: Forced phone home to verify Persistent entitlements are impractical given frequent rotation Open Loop: No forced phone home to verify Bare signatures on entitlements are unverifiable after rotation Unbounded fraud potential over time for any persistent entitlements that are not revoked at each rotation Undetectable fraud potential on unlinkable signatures (ZKPs) Specific: Detect, revoke, and reissue fraudulent entitlements Requires cryptographically verifiable append only log that binds issuances to key state at time of issuance
The bottom Ihe.ispfrée'solaing sucks. . gl s
T ek ere el No timely recoverability is the and your life'cangorfrom *Here | am : .
having a great:day out, moving over IT SeCUrlty equVaIent Of stone'getting-some mileage imfjust
me and therock,” to“Ohf***, oh
5**%, oh G %%, 0h~" BLAM. DEAD." —
Q o s Basas L
o S P M e L gf Jh
LA TR A A AT SO V74
SR TN AR TR e S
AR A A R N e T TRy d; N B B et FEA iR Re e s R AU s chess.comaledustsidefallespine Fespr mansalarchved pdt
Why signing infrastructure compromise is the hard problem Irresponsible to assume that compromise of the issuer signing infrastructure is impossible. Compromise enables verifiable forged issuances (impersonation fraud of the issuer, identity theft of every issuee), The risk of forgery persists throughout the lifespan of unanchored long-term issuances, even after the rotation of compromised keys. Unanchored forged issuances are not detectable (in general) and therefore preclude specific or timely recovery. Anon (ZKP) signatures are unlinkable to specific forged issuances, and hence, such forged issuances are undetectable by the issuer. Forged issuances that are unlinkable and hence undetectable by the issuer are not amenable to specific recovery. Given unanchored andjor unlinkable issuances, the only way to ensure full recovery from a likely or potential key ‘compromise is to revoke and reissue all issuances made with likely or potentially compromised keys. Prophylactic rotation bounds the time frame of potential compromise but at the cost of multiplying the cost of revoking and replacing ALL issuances at each rotation. Issuances to millions of citizens exponentially increases the ROI of attackers and, hence, the susceptibility of comprormise. Revoking and reissuing millions issuances due to likely or potential key compromise is impractical. Unanchored long-term non-post-quantum-safe signed issuances are vulnerable to the surprise quantum key compromise armageddon.
How to prevent subornment Horizontally scalable distributed signing infrastructure (keys at the edge) Nonrepudiable accountability of each and every employee (actions bound to actors) Detectability of any all forged issuances without reliance on cooperation/collusion with verifiers via anchored issuances tied to each employee Maker-checker-supervisor key management
Reidentfied —— Redanified [ Aggregater | Uier Agaregator s | Aggregaor | Pt 4 & ) Suried B / Issuarce Dedentifec / Deions / Pt / FTC Regaied o SOR0 S it / 7 S e N [— (v oo o — vertr £ % s = [ Exloter [[soie = o g R AR N~ S B ] o N lne \ B B e e s - s L A . Contrtal Gt Cotonual N Daonw Letkr Comocal o - o (o~ e Y o G e / N A S : o S — i v N — - O e = (R R \ activity / A disclosure L et L ed