Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for data engineers

Government Data Publication Preflight Agent

Datasets that are valid and safe to publish.

Government Data Publication Preflight Agent: what goes in, what the agent does and what you get

What it does

Open-data releases sometimes contain schema errors or hidden personal data. When a dataset is prepared, this agent validates it against the publication schema. It repairs structural problems, such as date formats or column names, in a working copy and revalidates. It then scans for restricted fields and personal data, paying close attention to free-text columns. If any disclosure risk appears, it stops release preparation and routes the issue to the disclosure officer instead of trying to clean it itself. The publishing team approves the release. Edge case: a comments column contains phone numbers, so the agent halts the release even though the schema checks pass.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN Dataset prepared 2 USES A TOOL Validate against the schema 3 CHECKS THE RESULT Does it pass structural checks? If not: make permitted repairs in the working copy. Backto step 2. 4 USES A TOOL Scan for restricted fields and personal data 5 CHECKS THE RESULT Is it free of disclosure risks? If not: stop and route the disclosure issue. Back tostep 4. 6 YOU APPROVE Publishing team approves release 7 RESULT Publication-ready dataset candidate
Read the steps as a list
  1. Dataset prepared
  2. Validate against the schema
  3. Does it pass structural checks?If not: make permitted repairs in the working copy. Back to step 2.
  4. Scan for restricted fields and personal data
  5. Is it free of disclosure risks?If not: stop and route the disclosure issue. Back to step 4.
  6. Publishing team approves releaseThe agent waits here for your OK.
  7. Publication-ready dataset candidate

How it decides

It repairs structural problems automatically and stops at any disclosure risk.

  • Repair structure only, never content.
  • Stop on any disclosure risk.
  • Free-text columns are always scanned.
  • Release needs the publishing team's approval.

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Publication schema to validate against
  • Personal data patterns to scan for (phone, email, address, ID numbers)
  • Structural repairs the agent may make
  • Who receives disclosure issues
  • Release schedule

What keeps you in control

It always asks you first

  • Public release
  • Disclosure decisions

Hard limits

  • Never publishes.

It stops when

  • Done: valid and safe.
  • Stopped: disclosure issue.

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensOn January 20 a dataset of 8,400 permit records is prepared. The schema check fails on 312 dates in the wrong format; the agent repairs them in the working copy and revalidates. The disclosure scan then finds phone numbers in 47 comment fields. That check fails, so it halts preparation and routes the issue. The disclosure officer removes the column, and the publishing team approves the release.

More agents for data engineers