Duval County Records Intelligence

The records already know. We do the reading.

Every morning, Duval County publishes who is behind on taxes, who was just sued over a mortgage, and whose estate has cleared probate. It arrives as raw clerk feeds nobody wants to parse. We parse it, match it to the parcel, grade it, and send it to you before the coffee is cold.

Tue–SatIn Your Inbox by Morning
$199Per Month, Cancel Anytime
CISSPCSSLP · CRISC

Think like an attacker. Write like an expert witness. Ship like an engineer.

Matched to Parcel
RE#, value, homestead
Graded
Best leads on top
The Flagship Data Product

The Duval Daily Hot Sheet

Newly filed distressed property records in Duval County, pulled from the Clerk every day, joined to county parcel data, graded, and delivered Tuesday through Saturday mornings. Lis pendens, tax liens, code liens, HOA judgments, recorded judgments, cleared probate estates, and tax deeds, in one file you can actually work from.

Filed yesterday, in hand today

The pipeline runs against the Clerk's feeds every recording day. You are working filings while they are still fresh, not a list somebody scraped last month.

Matched to the parcel

Every record carries the RE number, assessed value, homestead status, absentee owner flag, and mailing address. No second lookup before you can act.

Graded, not just dumped

Records are scored so the top of the sheet is the part worth working first. Two readable sheets plus three CSVs, ready for your CRM or your dialer.

County filings, blockchain ledgers, and mobile devices pose the same problem: the record exists, it is hard to reach, and it means nothing until somebody assembles it into a picture you can act on.

Shirley Singleton, Principal of Secure Data Consortium
Shirley Singleton
Principal · CISSP, CSSLP, CRISC
Who We Are

A builder and architect with adversarial depth

Secure Data Consortium is an independent records intelligence and security practice based in Jacksonville, Florida. It combines 19+ years building production systems across financial services, government, and security tooling with active adversarial research and digital forensics.

That matters for a data product more than it sounds like it should. The Hot Sheet pipeline is written and maintained by the same person who spent years deploying airgapped investigative data systems for government agencies. Record handling is the job here, not an afterthought bolted onto a spreadsheet.

Every engagement is project-shaped, handled start to finish by one experienced practitioner, and delivered as a written report.

More About the Practice
Investigative Practice

Where depth beats breadth

Alongside the data products, three deliberately narrow service lines. Each rewards experienced, hands-on expertise that has actually built and broken the systems in question.

Credentials & frameworks the work is grounded in

CISSPCSSLPCRISC OWASP LLM Top 10MITRE ATLASEU AI ActEx-Chainalysis
Why Work With Us

Experienced and hands-on, accountable end to end

Hands-On Expertise

Experienced, hands-on attention on every engagement, from first scope to final report.

Builder & Breaker

Production code and adversarial research together, a rare combination that this work demands.

Stated Limits

Data products and reports both say plainly what the records support and what they do not.

Project-Shaped

Defined scope, fixed timeline, fixed or capped fee. No open-ended retainers or staff-augmentation arrangements.

From the Lab

Independent research, published in the open

Public writeups are proof of work, the same methodology that goes into client engagements.

Reasoning Defense · 576 trials

Does reasoning make LLMs safer against prompt injection? Testing Qwen 3, DeepSeek-R1, and Gemma 4

A 576-trial controlled study across reasoning and non-reasoning open-weight models. Chain-of-thought reasoning lowers conventional injection rates, but attacks that target the reasoning step itself sidestep the defense, and reasoning costs roughly 15 to 20 times more tokens per call. Reasoning belongs alongside input and output controls, not instead of them.

Prompt Injection · 1,280 trials

Temperature is not a defense: indirect prompt injection across four open-weight LLMs

A 1,280-trial study of indirect prompt injection susceptibility across Llama 3.1 8B, Mistral 7B, Qwen 2.5 7B, and Qwen 2.5 Coder 7B. Key finding: temperature reduction is not a reliable defense; output-format constraint achieved 0 of 40 injection on the most-susceptible model tested.

Start a Conversation

Tell us what you're working on

Questions about the Hot Sheet, or a matter in one of the practice areas. Send a note below or schedule a 30-minute call.

Schedule a Call