<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en"><title>DataParade</title><id>https://dataparade.io/</id><updated>2026-10-14T00:00:00Z</updated><link rel="alternate" type="text/html" href="https://dataparade.io/"/><link rel="self" type="application/atom+xml" href="https://dataparade.io/posts/atom.xml"/><author><name>DataParade</name></author><subtitle>Articles from DataParade on preventing privacy and security problems through solid architecture and a real understanding of the system.</subtitle><entry><title>A Name Isn't an Identity</title><id>https://dataparade.io/posts/a-name-isnt-an-identity/</id><link rel="alternate" type="text/html" href="https://dataparade.io/posts/a-name-isnt-an-identity/"/><updated>2026-10-14T00:00:00Z</updated><published>2026-10-14T00:00:00Z</published><summary>A code scanner finds a hundred places where something called email appears. How many data items is that? Joining on the name alone dropped our grouping precision from 86% to 51%. The fix was to stop trusting names and join only on evidence the parser can show you.</summary><author><name>Ryan Porter</name></author><category term="Scanner"/><category term="Code Analysis"/><category term="Precision"/><category term="Parser"/></entry><entry><title>A Decision Model That Walks the Parse Tree</title><id>https://dataparade.io/posts/a-decision-model-that-walks-the-parse-tree/</id><link rel="alternate" type="text/html" href="https://dataparade.io/posts/a-decision-model-that-walks-the-parse-tree/"/><updated>2026-10-08T00:00:00Z</updated><published>2026-10-08T00:00:00Z</published><summary>Sorting data item mentions into data items is an agent task: look at a candidate group, read around it, decide. A frontier coding model does it well and costs too much to run on every scan. We built the same loop around Jev, gave it parse-tree tools instead of a file viewer, and it landed above the…</summary><author><name>Ryan Porter</name></author><category term="Jev"/><category term="Decision Models"/><category term="Parse Tree"/><category term="Cost Optimization"/></entry><entry><title>Your Coding Agent Already Knows Where the Password Went</title><id>https://dataparade.io/posts/your-coding-agent-already-knows-where-the-password-went/</id><link rel="alternate" type="text/html" href="https://dataparade.io/posts/your-coding-agent-already-knows-where-the-password-went/"/><updated>2026-10-02T00:00:00Z</updated><published>2026-10-02T00:00:00Z</published><summary>Every decision about how a system handles personal data now gets made in a conversation with a coding agent, and the transcript is sitting on disk. We wrote a tool that reads those transcripts. The next step is an interviewer that forms a hypothesis from them and asks you to confirm it, instead of…</summary><author><name>Ryan Porter</name></author><category term="BotSpy"/><category term="Coding Agents"/><category term="Session Traces"/><category term="Shift Left"/></entry><entry><title>We Labeled Six Repos with a Thousand Sub-Agents</title><id>https://dataparade.io/posts/we-labeled-six-repos-with-a-thousand-sub-agents/</id><link rel="alternate" type="text/html" href="https://dataparade.io/posts/we-labeled-six-repos-with-a-thousand-sub-agents/"/><updated>2026-09-29T00:00:00Z</updated><published>2026-09-29T00:00:00Z</published><summary>You cannot improve a scanner you cannot score, and nobody had labeled where the personal data lives in a real codebase. So we had three sizes of model label it, argue about it, and let the biggest one arbitrate. The labels are not perfect. They are a compass needle, and the first thing they found w…</summary><author><name>Ryan Porter</name></author><category term="Evaluation"/><category term="Sub-Agents"/><category term="Ground Truth"/><category term="LLM Evals"/></entry><entry><title>The Diagram Is a Projection</title><id>https://dataparade.io/posts/the-diagram-is-a-projection/</id><link rel="alternate" type="text/html" href="https://dataparade.io/posts/the-diagram-is-a-projection/"/><updated>2026-09-17T00:00:00Z</updated><published>2026-09-17T00:00:00Z</published><summary>A dataflow diagram is not the output of a scan. It is a query over a knowledge base, and the gaps in that knowledge base tell you exactly which questions to ask a person. We built an interviewer on that idea. It drew its first system-context diagram after one question. It also makes dumb decisions,…</summary><author><name>Ryan Porter</name></author><category term="OCSF"/><category term="Knowledge Base"/><category term="C4 Model"/><category term="Interview Engine"/></entry></feed>
