← All programs
COURSE / TECHNICAL AIAsk about availability
AI agents: permissions, data, and prompt injection
A technical program for building tool-connected agents with clear boundaries, least privilege, validation, and human oversight.
DURATIONFlexible program
FORMATIn person or online
Who it’s for
Developers, architects, and security teams building or assessing AI agents with access to data and tools.
Program
- 01 · Trust boundaries between instructions and data
- 02 · Identity, least privilege, and secret handling
- 03 · Validation, approvals, and action logs
- 04 · Prompt injection testing and risk review
What you’ll take away
- A threat model for an agent
- Controls to limit actions and access
- A pre-production testing checklist
PROGRAM / AJLEARN
WITH PURPOSE.ANGEL JOSE / TRAINING
WITH PURPOSE.ANGEL JOSE / TRAINING