7 papers published on arXiv

Qualixar

AI Agent Reliability Engineering

Research platform. Open source tools. Published papers.
Making AI agents trustworthy.

Explore ProductsRead Our Research
0
Published Papers
0
Open Source Tools
0+
Monthly Downloads
0
Academic Citations

Products

Open source tools for AI agent reliability

Qualixar OS

Qualixar OS

The Universal OS for AI Agents

Coming Soon

12 topologies, Forge AI auto-design, 24-tab dashboard, cost-quality-latency routing, judge pipeline. 2,831 tests.

$npx qualixar-os
SuperLocalMemory

SuperLocalMemory

Local-First AI Agent Memory

LIVE

Fisher-Rao geometry. 5-channel retrieval. Zero cloud dependency. EU AI Act compliant. 5,117+ npm downloads/month.

$npm install superlocalmemory
SLM Mesh

SLM Mesh

Peer-to-Peer Agent Communication

LIVE

Real-time P2P messaging between AI agent sessions. File locking. Shared state. 480 tests. MIT licensed.

$npm install slm-mesh
AgentAssay

AgentAssay

Token-Efficient Agent Testing

LIVE

78-100% cost reduction in agent evaluation. Stochastic 3-valued verdicts. 10 adapters. 7,605 trials.

$pip install agentassay
AgentAssert

AgentAssert

Behavioral Contracts for AI Agents

LIVE

ContractSpec DSL, drift detection, recovery re-prompting, compositional guarantees. 10 domains, 7 models.

$pip install agentassert
SkillFortify

SkillFortify

Agent Skill Security Analysis

LIVE

100% precision. 22 frameworks. 3 academic citations. DY-Skill formal model. ASBOM generation. 540-skill benchmark.

$pip install skillfortify

7 Published Papers

Peer-reviewed research on arXiv and Zenodo

2026

Qualixar OS: A Universal Operating System for AI Agent Orchestration

arXiv:2604.06392
2026

SuperLocalMemory V3.3: The Living Brain — Cognitive Memory for AI Agents

arXiv:2604.04514
2026

SuperLocalMemory V3: Information-Geometric Agent Memory

arXiv:2603.14588
2026

AgentAssay: Token-Efficient Agent Evaluation via Stochastic Verdicts

arXiv:2603.02601
2026

SkillFortify: Securing the AI Agent Skill Supply Chain

arXiv:2603.00195
2025

AgentAssert: Behavioral Contracts for AI Agent Compliance

arXiv:2602.22302
2026

SuperLocalMemory V2: Privacy-Preserving Multi-Agent Memory

arXiv:2603.02240

AI Agent Reliability Engineering

A new discipline at the intersection of testing, security, memory, and orchestration for AI agents. Six pillars. Seven papers. One platform.

01

Testing

Evaluate agent behavior with token-efficient benchmarks and adapter-based testing.

AgentAssay
02

Security

Detect skill injection, prompt leakage, and unsafe tool invocations before deployment.

SkillFortify
03

Memory

Local-first persistent memory with Fisher-Rao geometry and semantic retrieval.

SuperLocalMemory
04

Orchestration

12 topologies, lifecycle management, and a universal command protocol for agents.

Qualixar OS
05

Contracts

Formal behavioral contracts — preconditions, postconditions, and invariants for agents.

AgentAssert
06

Communication

Peer-to-peer agent messaging, shared state, and file locking across sessions.

SLM Mesh

Stay in the Loop

Weekly AI agent reliability insights, tool updates, and research highlights. No spam. Unsubscribe anytime.