Digital Commons Network^™

Hybrid Queue Management For Value-Based And Chronological Content, Joshua Chandiramani, Husain Wafaie, Danny Tom, Shreya Sawkar, Dan Padgett, Derek Reiersen

Defensive Publications Series

This disclosure describes a method and system for managing a prioritized queue that integrates content items from multiple sources, specifically balancing items assigned a value-based metric with items following a chronological delivery rule. By utilizing a linear data structure and a specialized sorting logic, the system ensures that high-value entries are prioritized for delivery while chronological entries maintain their relative order and predictable advancement through the queue. This dual-logic approach prevents the systematic deprioritization of items that lack a financial metric, thereby satisfying delivery expectations for diverse campaign types.

Keywords: Queue Management, Priority Sorting, Linked List, Mobile Content Distribution, …

Physically‑Aware Fault Modeling And Test Pattern Selection Using Multi‑Layer Perceptron Networks, Sandipan Sharma, Srinivas Vooka, Maheedhar Jalasutram, Pranav Murthy

Defensive Publications Series

Conventional semiconductor fault models often assume a linear correlation between fault counts and chip fallout, assigning equal weight to each fault node.  This approach fails to account for non‑linear outliers attributed to specific physical design features, leading to suboptimal fallout prediction and inefficient test pattern prioritization.  A method is disclosed that utilizes a multi‑layer perceptron network to generate physically aware fault models.  A fault feature matrix is constructed from physical design data that incorporates features (e.g., total net length, via density, proximity to power rails).  This matrix is processed by a fully connected neural network trained on production fallout data …

Hardware‑Interlocked Power Request Signaling For Secure Root‑Of‑Trust Operations, Rohit Sinha, Naved Kazi, Anchal Vijay

Defensive Publications Series

Traditional power management for secure memory operations often depends on external software to identify and request high‑voltage rails before sending commands to a secure entity. This reliance may create synchronization risks where memory programming might begin before power is stable, and involves non‑secure components inspecting secure directives, potentially exposing sensitive information. This publication describes a hardware‑interlocked handshake mechanism between a Root‑of‑Trust and a Power Manager to manage power rails on‑demand. The Root‑of‑Trust independently identifies the request for high‑voltage power and asserts a hardware request signal. The Power Manager receives a hardware interrupt, enables the indicated power rail, and returns a …

Bypassing Distributed Access Control For Subsystem‑Level Debugging, Rohit Sinha, Sumit Jha, Mayank Tutwani

Defensive Publications Series

Distributed access control in modern heterogeneous architectures often leads to debugging complexities during software development. Security enforcement logic is typically embedded across multiple initiators and targets, creating a wide search space when failures occur. It is often difficult to distinguish between functional software bugs and access control misconfigurations, as both may result in similar system behavior or silent failures. A system and method are disclosed for the selective relaxation of access control policies through a hardware‑assisted debug controller. The debug controller may provide granular enable signals to individual subsystems. Each subsystem may include a control status register that, when qualified …

A Hybrid Cryptographic Encoding System Combining Aes-256- Gcm Authenticated Encryption, X25519 Forward Secrecy, Shamir Threshold Key Management With Biometric Access Gating, And A Pluggable Steganographic Output Layer, Naveen Rajan

Defensive Publications Series

A hybrid cryptographic encoding system is disclosed in which plaintext is encrypted using AES256-GCM keyed via an X25519 ephemeral-static Diffie-Hellman key agreement, optionally protected at rest by a Shamir K-of-N secret-sharing scheme whose shares are individually gated by biometric fuzzy extractors, and the resulting authenticated ciphertext is rendered through a pluggable steganographic output layer (key-permuted Morse, Base64, or linguistic steganography). The combination delivers confidentiality, integrity, forward secrecy, threshold multi-party access control, and format-agnostic camouflage in a single interoperable pipeline.

Codex Oracle Architecture (Coa): Signal‑Driven, Meta‑Orchestrated Multi‑Agent Coordination For Autonomous Code Generation, Johny Manuel-Devadoss

Defensive Publications Series

This disclosure describes a novel orchestration architecture for multi‑agent code generation systems, referred to as the Codex Oracle Architecture (COA). COA introduces a signal‑driven coordination layer, a meta‑orchestrator (“Oracle Layer”), and a structured conflict‑resolution mechanism that enables large‑scale, adaptive, and emergent collaboration among heterogeneous AI agents. Unlike existing multi‑agent frameworks that rely on static workflows, deterministic message passing, or hierarchical control, COA enables dynamic agent activation, probabilistic decision making, and context‑aware collaboration. The architecture is designed to support autonomous code generation, refactoring, debugging, and long‑term software evolution in complex environments.

Human-In-The-Loop Workflow For Ai-Assisted Preliminary Bug Investigation, Mira Leung

Defensive Publications Series

A disclosed technology can address challenges in managing new software bugs when engineering capacity is limited, a situation which can be impacted by time-consuming initial investigations. The technology can provide a structured, human-in-the-loop workflow where a bug reporter may use an artificial intelligence (AI) assistant, such as a large language model, to perform a preliminary analysis at the point of bug discovery. This process, which can be guided by structured prompts, may generate an enriched bug report containing analysis from the AI, including, for example, potential root causes, code pointers, and ownership suggestions. By shifting a portion of the initial …

Ai-Driven Pipeline For Automated Prompt Refinement Using Evaluation Feedback, Cristina Elena Budurean, Quinn Halpin, Ibrahim Gunay

Defensive Publications Series

Conventional methods for improving generative artificial intelligence (AI) prompts can rely on manual analysis of evaluation feedback, a process that may be time-consuming, subjective, and may not scale effectively. A data-driven pipeline is described that can facilitate automated prompt refinement. The system may operate in a feedback loop, ingesting evaluation data, for example, rater scores and qualitative comments. The system can then use a large language model to perform a multi-stage analysis that may include sanitizing data, diagnosing response failures, extracting successful patterns, and resolving potentially contradictory feedback to generate a revised prompt. This approach can support a more scalable, …

Active Permeability Control Of A Magnetic Shield For Haptic Actuators, Vincent Chung, Vander Fang, Jianxun Wang, Matilda Lai, Chuck Tally

Defensive Publications Series

Magnetic interference between haptic actuators and sensitive magnetic sensors can present a challenge in compact electronic devices, such as smartphones or wearable devices. While a static high-permeability shield may contain stray magnetic fields from an inactive actuator, it could also impede the magnetic flux during haptic operation, potentially reducing actuator efficiency. Systems and methods can address this by actively controlling the shield's permeability. For example, a shield made from a high-permeability ferromagnetic material can at least partially enclose the haptic actuator. A permeability-modulation circuit and an associated drive coil can be used to selectively drive the shield material into magnetic …

System For Real-Time User Compositing Into Media Streams Using Generative Ai Conditioned On Extracted Scene Context, Dev Daftari, Akash Verma

Defensive Publications Series

A system is described for compositing a user's image into media streams protected by digital rights management (DRM). The system may operate via a multi-stage pipeline, which can begin with context extraction on a user device, such as a smartphone or smart television. Computer vision models may analyze a media stream to create a context packet detailing information such as scene lighting, spatial depth, and artistic style. This packet can then be used to condition a generative artificial intelligence (AI) model, which may synthesize the user's likeness into the scene with corresponding visual characteristics. A resulting composite can be rendered …

Border Gateway Protocol Prefix Limit Negotiation Or Prefix Limit Advertisement, N/A

Defensive Publications Series

The technology described in this paper relates to negotiating and advertising border gateway protocol (BGP) prefix limits between peering network devices. By exchanging configured prefix limit values during BGP session establishment, sending and receiving devices can establish an operational prefix limit to avoid unexpected session terminations. The lowest configured limit can serve as the operational limit, or a receiving device can advertise its limit to a sending device to cap advertised routes. Subsequently, when limits are reached, newer or random prefixes for sessions, if a session is restarted or cleared, can be blocked rather than being terminated.

Semantic Discovery And Composite Ranking Of Skill Instruction Files In An Ai Agent Marketplace, Aaron Burton

Defensive Publications Series

A method for discovering and ranking skill instruction files published to an AI agent marketplace using a composite scoring function applied to a vector index. When a querying agent submits a natural language query expressing a desired capability outcome, the system computes a composite relevance score for each indexed skill by combining embedding cosine similarity between the query and the skill description, declared constraint compatibility between the skill and the querying agent's operational profile, and an optional historical quality signal derived from prior benchmark completions. The ranking layer operates entirely upstream of any licensing, payment, or entitlement system and returns …

Structured Telemetry Span Schema For Ai Agent Skill Instruction Execution, Aaron Burton

Defensive Publications Series

A method for emitting structured observability telemetry during AI agent execution by keying each telemetry span to entries within a skill instruction file. Each span record carries a span identifier, a parent span identifier for trace hierarchy, a reference to the originating skill file, a reference to the specific capability entry being exercised, an optional reference to any constraint that was evaluated during the span, a hash of the pre-execution context, a hash of the post-execution output, elapsed duration, a list of tool invocations made during the span, and an optional score field that records a numeric value produced by …

Cross-Provider Skill File Portability Shim For Behavioral Constraint Preservation Across Ai Tool Frameworks, Aaron Burton

Defensive Publications Series

A method for translating a canonical AI skill instruction file into provider-specific tool definitions for multiple target AI frameworks while preserving behavioral constraint metadata through explicit mapping rules. The skill instruction file is authored in a single canonical format containing structured capability entries with execution parameters and behavioral constraints. A translation layer ingests this file and emits provider-specific adapter definitions for OpenAI function calling JSON schema, Anthropic Claude tool-use schema, LangChain Tool class Python stub, and Model Context Protocol capability manifest JSON. Constraints that have no direct equivalent in the target framework are emitted as a standardized annotation block rather …

Skill Instruction File Dependency Resolution And Version Pinning For Ai Agent Runtimes, Aaron Burton

Defensive Publications Series

A method for resolving cross-file dependencies between AI agent skill instruction files using semver-style version range constraints declared within each file, analogous to dependency resolution in software package managers such as npm, pip, and Cargo. Before agent invocation, a resolver traverses the dependency graph formed by inter-skill "requires" declarations, computes a consistent version assignment satisfying all range constraints across the entire graph, detects version conflicts between incompatible requirements, and either selects a valid resolution or emits a structured conflict report identifying the incompatible pair. A sidecar lockfile is produced at the completion of resolution, recording the exact resolved version of …

Multi-Agent Skill File Negotiation Protocol For Stateless Capability-Based Workflow Assignment, Aaron Burton

Defensive Publications Series

A method for coordinating task assignment across multiple AI agents in a multi-agent workflow by having each agent advertise its capabilities and constraint parameters at the start of each session, with a stateless message broker collecting those advertisements and selecting an assignment that covers the full capability set of the requested workflow at the lowest aggregate constraint cost. Each agent re-advertises per session, eliminating dependence on a persistent central skill registry. Tie-breaking uses historical completion latency self-reported by each agent. The method enables capability-based task distribution across heterogeneous AI agents without requiring any agent to hold global knowledge of its …

Arachne Aero-Skin, Michael Victor Caldwell Mr.

Defensive Publications Series

The Arachne Aero-Skin is a passive aerodynamic surface film designed to reduce aircraft skin-friction drag through microstructured riblet geometry aligned with local airflow. The system is intended for external application on non-structural aircraft surfaces and is designed for manufacturability, durability, and measurable performance under realistic contamination, cleaning, and environmental conditions.

This Phase 1 system excludes active flow control, propulsion modification, plasma systems, or engine integration.

The Virtual Content Viewing Device That Consists Of 8 Projection Configurations That Are Spatially Configured In A Curved And An Adjacent Manner To Provide A Greater Virtual Field Of View That Includes The Far Peripheral Areas., Punarjeewa Abeysekera

Defensive Publications Series

This paper describes a virtual content viewing device that consists of 8 projection configurations that are spatially configured in a curved and an adjacent manner to provide a greater virtual field of view that includes the far peripheral areas. The entire virtual content viewing device altogether consists of 8 identical projection configurations. And the entire virtual content viewing device will consist of a left projection system and a right projection system. The left projection system will contain 4 identical projection configurations. And the right projection system will contain 4 identical projection configurations. The containing 8 identical projection configurations will have …

Privacy-Preserving Cross-Customer Threat Intelligence Platform With Embedded Ml, Predictive Attack Trajectory, And Multi-Agent Trust Management For Web, Mobile, And Ai Agent Protection, Duncan Ndegwa Ndungu

Defensive Publications Series

This publication discloses twelve platform-scale security intelligence inventions for web, mobile, and AI agent protection. Core platform inventions (B1–B4) include: privacy-preserving cross-customer threat intelligence via a two-layer alias architecture with k-anonymity; SDK-embedded quantized ML inference (<100KB, INT8, sub-5ms) with knowledge distillation and signed over-the-air updates; consent-scoped alias lifecycle management for automatic GDPR/CCPA compliance; and multi-hop AI agent trust chains with configurable attenuation and cascading revocation. Advanced security infrastructure (B5–B10) covers IDOR detection via per-alias resource access baselines, runtime dependency integrity verification via Ed25519-signed manifests, AI agent output exfiltration detection via semantic and entropy analysis, multi-level TLS certificate pinning with safe rotation and air-gapped fallback, constant-time authentication response orchestration, and shadow agent detection via distributed registry and behavioral fingerprinting. Predictive infrastructure (B11–B12) includes DNS hijacking detection via multi-path resolution consensus with challenge-response verification, and a Bayesian predictive attack trajectory engine with pre-emptive countermeasure deployment and outcome feedback learning.

Customer-Facing Api Key Aliasing System With Policy-Bound Lifecycle, Zero-Downtime Rotation, Third-Party Data Payload Protection, And Closed-Loop Revocation For Web, Mobile, And Agent Credential And Data Security, Duncan Ndegwa Ndungu

Defensive Publications Series

This publication discloses a customer-facing API key aliasing system (A19) and a companion third-party data payload protection system (A19B). A19 provides alias indirection for API credentials across web, mobile, CI/CD, AI agent, and third-party integration environments — with policy-bound alias issuance, zero-downtime dual-alias rotation, device attestation binding for mobile (Apple DeviceCheck, Google Play Integrity), selective alias revocation, anomaly monitoring, and closed-loop revocation. A19B adds three layers of application-layer security above TLS: multi-identity outbound payload signing binding five identity dimensions (application, user, browser, device, alias) into a single HMAC-SHA256 attestation; selective AES-256-GCM field-level encryption with automatic PII detection and field-path AAD …

Comprehensive Pre/Post-Authentication Attack Detection And Mitigation Framework For Web, Mobile, And Ai Agent Applications, Duncan Ndegwa Ndungu

Defensive Publications Series

This publication discloses fourteen specialized attack detection modules operating within an SDK event observation pipeline at zero incremental infrastructure cost, covering both pre-authentication and post-authentication phases across web, mobile, and AI agent applications. Pre-authentication modules (A5–A10) detect password spraying via cross-account fingerprint correlation, mass registration abuse, AI agent tool-call sequence violations and infinite loops, OAuth redirect hijacking, and SDK configuration tampering. Post-authentication modules (A11–A18) detect MFA fatigue attacks, session fixation, business logic abuse via learned transaction flow graphs, refresh token theft via device fingerprint binding, indirect prompt injection via pre-ingestion content scanning, feedback poisoning, and clock manipulation via dual-clock comparison. …

Adaptive Behavioral Security System With Real-Time Closed-Loop Response For Web, Mobile, Api, And Ai Agent Protection, Duncan Ndegwa Ndungu

Defensive Publications Series

This publication discloses six interrelated security inventions implemented in the DevFortress SDK v4.8.0: (1) Credential Aliasing — replacing real authentication credentials with cryptographically random aliases (supporting 10 generation methods including CSPRNG, HKDF-SHA256, BLAKE3, and FPE/FF3-1) so that a complete breach of the monitoring service yields zero valid credentials; (2) a Three-Mode Closed-Loop Response Engine achieving sub-10ms automated threat remediation via external, internal, and hybrid modes; (3) a Behavioral Baseline Engine using adaptive EWMA across eight feature dimensions for novel attack detection; (4) Cross-Alias Correlation using Locality-Sensitive Hashing for distributed attack identification and simultaneous cluster revocation; (5) Graduated Rate Limiting with …

Temporal Blind Spot Failure In Serverless Cloud Systems: How Ai Agents Cause Cloud Disasters, Aaron Burton

Defensive Publications Series

This disclosure defines and documents Temporal Blind Spot Failure, a recurring agentic AI failure mode in which autonomous AI agents operating inside serverless cloud systems take actions that are locally rational but catastrophic across complex cloud environments. Temporal Blind Spot Failure is distinct from prompt injection, memory poisoning, and output hallucination. It occurs because the agent decides one step at a time and cannot see how its choices compound across the systems those choices touch. This disclosure establishes Temporal Blind Spot Failure as a named category of agentic AI failure modes, describes how it shows up across the model, the …

Configurable Document Requirement Checklist Engine For Real Estate Transaction Types, Aaron Burton

Defensive Publications Series

A configurable checklist engine for real estate document collection that maintains per-transaction-type requirement definitions specifying required documents, acceptable file formats, expiration rules, and submission deadlines. The engine tracks completion percentage across all parties in a transaction by comparing submitted and verified documents against the active checklist. Transaction types include residential purchase, commercial lease, refinance, and 1031 exchange, each with distinct document requirement sets that administrators can modify without code changes.

Multi-Party Transaction Role Mapping With Configurable Document Visibility And Distribution Permissions For Real Estate Workflows, Aaron Burton

Defensive Publications Series

A system for defining and managing role templates specific to real estate transactions, where each transaction type maps predefined participant roles to specific document visibility and distribution permissions. Roles include buyer agent, listing agent, lender, title officer, escrow agent, inspector, and appraiser. Each role template specifies which document categories a participant in that role can view, which documents they can receive via distribution, and whether they can initiate document requests. Role templates are configurable per subscribing organization and per transaction type, allowing the same platform to serve brokerages, lenders, and title companies with different permission models.

Dual-Party Document Request Routing And Fulfillment Tracking For Multi-Sided Real Estate Transactions, Aaron Burton

Defensive Publications Series

A method for managing document collection in dual-party real estate transactions where both buyer-side and seller-side participants can issue document requests through a shared platform. The platform routes each request to the appropriate submitter based on that submitter's transaction role, tracks document fulfillment independently per party side, and maintains separate completion status indicators for the buyer side and seller side of the transaction. The method enables transaction coordinators and administrators to monitor the overall document readiness of a transaction from a single interface without commingling buyer-side and seller-side document obligations.

Deliberative Constraint-Field Multi-Agent Systems (Dcf-Mas): A Framework For Fate-Constrained, Narrative-Consistent Autonomous Workspaces, Johny Manuel-Devadoss

Defensive Publications Series

Contemporary multi-agent systems are predominantly designed around optimization, task decomposition, and explicit coordination protocols. While effective in structured environments, these paradigms struggle with long-horizon reasoning, irreversibility, and systemic constraints that are not directly observable. This paper introduces Deliberative Constraint-Field Multi-Agent Systems (DCF-MAS), a novel framework that reconceptualizes multi-agent coordination as a process of collective deliberation operating within latent constraint fields and irreversible state transitions. DCF-MAS departs from optimization-centric orchestration by introducing three foundational mechanisms: Shared Deliberative Substrates, Latent Constraint Fields, and Irreversible State Transition Engines. These mechanisms enable agents to reason collectively, adapt to hidden systemic limitations, and account for …

Siege-State Multi-Agent Workspace Systems (Ss-Maws): Constraint-Driven, Encirclement-Based Ai Coordination For Persistent Adversarial Environments, Johny Manuel-Devadoss

Defensive Publications Series

Existing multi-agent workspace managers emphasize task decomposition, cooperative execution, and centralized or semi-distributed orchestration. While effective in stable environments, these systems lack robustness under adversarial conditions, prolonged uncertainty, and resource contention. This paper introduces Siege-State Multi-Agent Workspace Systems (SS-MAWS), a novel framework inspired by siege warfare doctrine, in which progress is achieved not through direct execution but through gradual constraint imposition, resource isolation, and systemic pressure.

SS-MAWS reconceptualizes multi-agent coordination as a process of encirclement, containment, and controlled resolution, rather than linear task completion. The framework introduces three foundational constructs: (i) Encirclement Fields, which dynamically constrain problem …

Crusade Doctrine-Inspired Multi-Agent Workspace Governance (Cd-Mawg): A Futuristic Framework For Post-Orchestration Ai Workspaces, Johny Manuel-Devadoss

Defensive Publications Series

Contemporary multi-agent workspace managers rely heavily on centralized orchestration, graph-based execution models, and static role assignment. While effective for bounded workflows, these paradigms exhibit structural limitations when extended to long-horizon, adaptive, and adversarial environments. This paper introduces Crusade Doctrine–Inspired Multi-Agent Workspace Governance (CD-MAWG), a novel architectural framework that reconceptualizes multi-agent coordination through the lens of crusade-era military organization, logistics, and distributed authority.

Rather than treating computation as a sequence of tasks, CD-MAWG models it as a persistent campaign sustained by evolving agent societies. Core innovations include relic-based authority propagation, campaign-state memory structures, emergent agent castes, and dual-constraint scheduling grounded …

System For An Interactive Ambient Character Using Mmwave Radar For Presence Detection And Acoustic Event Recognition, Hanumanth Godari

Defensive Publications Series

A system can use an on-screen digital character powered by an artificial intelligence framework to provide an interactive ambient display. The system may integrate a millimeter-wave radar sensor for camera-less presence detection and a far-field microphone array for acoustic event recognition of ambient household sounds. Based on data from these sensors, the digital character could, for example, acknowledge a user entering a room or react to specific sounds, such as music or a doorbell. This allows a display's idle screen on a computing device (e.g., a smart display, television, or tablet) to function as a dynamic, context-aware interface. The interface …