Agentic Security Pedia
What to ask about every AI agent
Plain answers to the questions security, compliance and engineering leaders ask about AI agents. One question per article, answered in its first paragraph, with sources you can check and what auditors ask.
The basics
How agents get attacked
Controls
Evidence and compliance
One question each
Every article answers one question, and the answer comes first, in under fifty words.
Sources you can check
Every fact links to where it came from: the standard, the advisory or the original research.
Reviewed, and dated
Each article names its author and the date it was last reviewed, and maps to the frameworks auditors use.
More from GIRD
Incident dossier
38 coding-agent incidents, sourced
What went wrong, and what would have stopped it.
Case study19 days on one company's laptops
Four keys left inside ordinary agent requests. None looked like an incident.
For your engineersBlindspot
An open-source, read-only check for the AI keys and agents on one computer.