The research library / Six connected fields

Good questions
deserve a closer look.

Working questions, experiments, and methods behind the software. Browse by field or follow a topic into its context and related work.

17 of 47 research topics · AI systems

01

ALX Semantic Compiler

AI systems

Source-preserving semantic compilation, execution contracts, and bounded repair with explicit evaluation limits.

Active
02

Protocol Compiler Framework

AI systems

A repeatable method for compiling research and operating rules into agent-usable protocols.

Canonical
03

ALX LoRA Evaluation Track

AI systems

Evaluation plans for whether specialized training can preserve ALX structure and reversibility.

Current
04

Home Base Agent OS

AI systems

Project registry, memory, governance, source ledgers, and recovery paths for long-running agent work.

Active
05

Task-Aware Context Rollover & Recovery v1.1

AI systems

A protocol for preserving intent, evidence, and next actions when context windows or sessions change.

Canonical
06

AI Fundamentals & Markdown Transformer Compiler

AI systems

A structured language and transformation lane for turning raw knowledge into agent-readable artifacts.

Current
07

Forge Scheduler / FOS

AI systems

Research into scheduling, bounded execution, and operating-system-like coordination for agent workflows.

Prototype
08

Browser Agent Consultation Prompt Pack

AI systems

Reusable consultation patterns for browser agents that separate observation, judgment, and action.

Current
09

Compute Broker-Dealer

AI systems

A prompt and architecture track for allocating compute, models, and work according to task value and constraints.

Prototype
10

MCP-First UI & Agent Pipelines

AI systems

Interface patterns where tools, permissions, evidence, and agent actions remain explicit from protocol to UI.

Active
11

Framework Selection Protocol Forge

AI systems

Evidence-led framework selection, including a current Svelte agentic-coding research branch.

Series
12

Shared Skills Library

AI systems

Portable skills and operating instructions designed to transfer across tools and model providers.

Active
13

Neuron Protocols

AI systems

Provider-independent orchestration, scoped retrieval, tools, memory, and surface-aware interaction.

Active
14

Laya and Jev Decision Probe

AI systems

A small typed-decision comparison and a test of prompt routing for arithmetic, with the limits kept visible.

Prototype
15

Persistent Agent Memory

AI systems

Portable project notes and explicit recovery state across assistants and sessions.

Active
16

Deterministic Prompt Routing

AI systems

Route tasks to operating protocols without making provider identity part of the decision.

Active
17

Scope & Execution Contracts

AI systems

Define observable acceptance criteria, constraints, and named authority before a broad build.

Canonical

This library presents public summaries of research programs and methods. A topic’s presence here does not imply a published paper or a validated result; individual briefs distinguish the available summary from the work still under investigation.

Stupid PeopleCheck the game’s Menu for save status.