About TKOResearch
Software, security
& lab work.
Our work covers AI Security, Cybersecurity, and Device & Firmware Assessments: from agent permissions and system architecture to components, firmware and materials.
Based in Alexandria, Virginia, TKOResearch builds software and takes on security engineering work with technical teams, company leadership and private clients.
What we're building.
TAREOps.
TAREOps is an AI-native, next-generation Laboratory Information Management System (LIMS) with an integrated Digital Lab Notebook and Digital Lab Assistant. It is designed to connect data ingestion, analysis and controlled AI actuation while keeping the sample, method and record connected to the work.
A lab result needs context: the data and method that produced it, changes to the record and the actions that were authorized. Those details shape how we build the notebook, assistant and operational controls.
TARE stands for Technical Analysis & Research Engine. Lab records and sample tracking are available through design-partner access. AI-assisted documentation is still in research and development, and deep instrument ingestion is on the roadmap. Actuation is experimental and requires workflow-specific validation. The maturity guide separates current capabilities from planned work.
Read about TAREOpsResearch & tools.
Our published work includes guides to agent permissions and logging, planning tools, and source code you can inspect.
- What an AI Agent Audit Trail Needs to Record
What to log when an agent requests access or takes an action.
- AI Risk Briefing Builder
A browser tool for drafting an AI risk report for a board meeting.
- GHS Label Builder
Prepare chemical labels and export them for printing.
Bring us the system and the question.
An engagement might test an AI agent's access to company data, review a product's security architecture, or examine a device whose components differ from its documentation. The service depends on the question you need answered.
AI Security
Find out how your AI systems handle sensitive data and actions. We assess existing workflows, build controlled integrations and help resolve security findings.
Cybersecurity
Get a technical review of your architecture, security controls or vendor claims. We help engineering teams, leadership, counsel and MSPs understand what needs attention.
Device & Firmware Assessments
Examine how a device is built, the components it contains and the firmware it runs. OASIS lab methods help answer questions about its condition, materials and behavior.
The OASIS lab supports our device work with stereo microscopy, electrical bench measurements, thermal imaging, FTIR spectroscopy and wet-lab methods. We agree which methods can answer the question before accepting a device or sample.
To discuss a project, send a short description of the system and the decision you need to make. We'll work out the access, methods and deliverables together before the work starts.
Background
Founded by Kevin O'Connor.
Founder & Principal Engineer
Kevin worked on CrowdStrike's research team and led threat research at Adlumin and N-able. At Adlumin, he led the team behind the published analysis of PowerDrop, a PowerShell backdoor found at a U.S. aerospace contractor.
His current research examines AI-agent permissions, tool access and the records needed to understand an agent's actions. He also writes at KevinBytes and publishes code on GitHub.