Learn

How large language models actually work with case materials, and where that breaks down.

The anatomy of a disciplined legal prompt — role, boundary, calibration, refusal, and review.

Reliability, absence, and provenance — reading a record the way Protocols 0.3 and 0.4 require.

Applying the method across cross-examination, voir dire, opening, and closing.

Why self-audit isn't verification, and what a real Protocol 0.7 pass looks like.

Confidentiality, disclosure, and professional-responsibility questions specific to AI-assisted defense work.