SourceReady field notes

Document security for everyday AI work.

Practical guidance on prompt injection, document structure and safer ways to bring outside files into AI projects.

Understand how instructions inside external content can influence an AI system—and where to interrupt that path.

Look beyond the visible page at the document structures, privacy boundaries and scan coverage that matter before AI use.

Practical routines for reviewing outside documents before they become project context, evidence or action.

Latest guidance

Start at the document boundary.

Learn how instructions inside external documents can influence AI, why the impact varies, and where a document scan fits.

Follow the path from a document’s hidden content to an AI model—and see where a scan can interrupt that path.

A PDF can contain more than its rendered pages. Learn which machine-readable surfaces matter before using one with AI.

Understand browser-local document scanning, which work can happen on your device, and which privacy details still deserve attention.

Use this practical checklist to review, scan and route outside documents before they become part of an AI project.

Build a safer research workflow by scanning external documents, preserving provenance and reviewing AI-generated conclusions.