IP Library Granted Patent US 12695750
Granted Patent B1
US 12695750 · App. 19/387,313 · Granted Jul 28, 2026

System and method for proxy-based credential custody with distributed validation

Inventor: Michael Pak (Portland, OR)
H04L63/0884G06F21/57H04L9/085H04L9/3278H04L63/0281
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12695750
App. No.
19/387,313
Granted
Jul 28, 2026
Kind
B1
Abstract

The present disclosure provides a distributed credential custody system that fundamentally separates credential possession from credential usage. The system eliminates the primary vulnerability of conventional API security by storing actual access credentials within a secure credential management system while providing clients only with opaque credential handles that cannot be independently used for authentication. Multiple independent authenticators perform distributed validation using hardware-anchored mechanisms, with each authenticator contributing weighted partial decryptions based on their assurance levels. The credential management system reconstructs plaintext credentials only after achieving threshold consensus from authenticators, then provides reconstructed credentials to a proxy. This architecture prevents credential theft from client systems while maintaining compatibility with existing authentication protocols, enabling organizations to implement zero-trust principles without requiring modifications to client applications or backend services. The distributed trust model ensures no single component can independently access complete credentials, providing enhanced security through cryptographic isolation and policy-driven authentication that adapts to real-time risk assessment.

Claims (38)

1 . A system for secure Application Programming Interface (API) access control, the system comprising:

a proxy configured to receive a resource access request from a client, the resource access request including a credential reference;

a credential management system (CMS) comprising one or more processors and memory storing instructions that, when executed by the one or more processors, cause the CMS to manage lifecycle of an access credential corresponding to the credential reference and coordinate with one or more authenticators to validate the client according to a credential policy;

the one or more authenticators, each authenticator comprising one or more processors and memory storing instructions that, when executed by the one or more processors, cause the authenticator to independently evaluate the client using at least one of passive validation or active challenge-response mechanisms and perform threshold decryption of the access credential upon successful validation of the client;

the CMS further configured to receive decrypted parts of the access credential and combine the decrypted parts into a cohesive decrypted access credential;

the CMS further configured to transmit the cohesive decrypted access credential to the proxy; and

the proxy further configured to provide the cohesive decrypted access credential to a target resource,

wherein the proxy operates as an intermediary that maintains exclusive access to the access credential during an injection process, and wherein the proxy is configured to provide the cohesive decrypted access credential to the target resource without transmitting the access credential to the client.

2 . The system of claim 1 , wherein the credential reference is an opaque identifier that lacks cryptographic value and cannot be independently used for API access.

3 . The system of claim 1 , wherein the active challenge-response mechanisms comprise proof of possession of a challenge-response pair generated by at least one of:

a Physical Unclonable Function (PUF) embedded in the client or a designated PUF response system; or

an immutable root-of-trust embedded in the client or a designated immutable root-of-trust response system.

4 . The system of claim 3 , wherein the Physical Unclonable Function produces a device-unique response derived from inherent physical variations in hardware of the client.

5 . The system of claim 1 , wherein the passive validation comprises validating the client using telemetry-based attributes including at least one of IP address, time of day, call frequency, application protocol, client software version, or configuration.

6 . The system of claim 1 , wherein:

the credential policy defines throttling enforcement parameters that limit data volume access for the client; and

the proxy is further configured to measure volume of response data for throttling enforcement.

7 . The system of claim 1 , wherein the CMS is further configured to rotate the credential reference upon expiration or renewal of the access credential.

8 . The system of claim 7 , wherein:

the CMS maintains audit logs of credential reference usage and validation events; and

failed client validations cause automatic revocation of the credential reference.

9 . The system of claim 1 , wherein each authenticator is assigned an integer weight representing its assurance level, and the CMS determines satisfaction of a threshold policy when a cumulative assurance weight greater than or equal to a predetermined threshold is achieved.

10 . The system of claim 9 , wherein the threshold decryption requires combining partial decryptions from at least t authenticators out of n available authenticators, where t and n are configurable based on the credential policy.

11 . The system of claim 1 , wherein the access credential is governed by an attribute-based access control policy specifying contextual attributes including at least one of identity, device integrity, geolocation, IP reputation score, behavioral history, and environmental and client telemetry.

12 . The system of claim 1 , wherein one or more authenticators are hardware-anchored using at least one of a Physical Unclonable Function, Trusted Platform Module, immutable root-of-trust, or Trusted Execution Environment.

13 . The system of claim 1 , wherein the CMS dynamically adjusts authentication requirements based on real-time risk indicators including at least one of device drift, network anomalies, or behavioral deviation.

14 . The system of claim 1 , wherein:

the access credential remains encrypted at rest within the CMS; and

the access credential is decrypted only transiently within secure memory of a protected execution boundary during the combine time by the CMS after successfully completing the threshold decryption process.

15 . The system of claim 14 , wherein the protected execution boundary comprises at least one of a Trusted Execution Environment or a Hardware Security Module.

16 . The system of claim 1 , wherein:

all authentication approvals and decryption events are cryptographically signed and hash-chained to form a tamper-evident audit log; and

the CMS maintains provenance tracking that records which authenticators approved access and which key shares were used for each credential reconstruction.

17 . The system of claim 1 , wherein the proxy presents the cohesive decrypted access credential to the target resource in standard protocol formats compatible with existing backend services without requiring modifications to the target resource.

18 . The system of claim 1 , wherein:

the system remains functional despite failure or unavailability of authenticators whose combined weight falls below the threshold requirement, where each authenticator has a weight greater than or equal to 1 and validation succeeds when the total weight of available authenticators is greater than or equal to threshold t; and

operational control of the CMS, proxy, and authenticators is distributed among distinct administrative domains to prevent single points of control.

19 . The system of claim 1 , wherein one or more authenticators include add-on authenticators, including user-developed and third-party authenticators, that are configured to be registered and executed as part of the client authentication process.