WritingTechnical notesFailure modes

Blog

Technical notes on AI agents, delegated authorization, runtime policy enforcement, and audit-grade receipts.

FoundationsIAMPolicyReceiptsVoice

Latest posts

Sort:Newest

FoundationsDec 2025

The Agent Trust Stack: Delegation → Enforcement → Receipts

A practical model for making autonomous agents safe in production systems, grounded in capability delegation, runtime policy, and audit-grade receipts.

Read →

IAMSep 2025

Authentication vs Authorization for Agents

Why service-to-service OAuth is not "an agent acting for a human," and how to model scoped, revocable authority with step-up approvals.

Read →

AuditNov 2025

Signed Action Receipts: Tamper-Evident Audit for Agent Actions

What receipts are, why logs are not enough, and how to design a signed receipt format that partners can verify independently.

Read →

VoiceJan 2026

Voice Deliverability Engineering: What Actually Burns Numbers

A technical breakdown of outbound calling failure modes, what you can deterministically control, and which metrics correlate with blocking risk.

Read →