Posts
20 posts
The Agent Memory Trilemma
The Tool Layer as the Agent Bottleneck
From Token Alignment to Representation-Level Defense
The Confused Deputy and Agent Identity
Embedding Inversion Attacks
RAG Datastore Membership Inference
LLM Agent Memory Poisoning
LLM Jailbreaks and Automated Red Teaming
MCP Tool Poisoning
LLM Privacy Defenses
LLM Privacy Attacks
Output Guardrails, Observability, and the Right to Erasure
Egress Gateway Defense
Defenses at the Agent, RAG, and MCP Layers
Input-Side PII Anonymization and Reversible Mapping
How Personal Data Leaks, and What Data Governance Has to Cover
Exposing a Home Server Safely with Cloudflare Tunnel
JWT Authentication and TOTP 2FA
TLS and HTTPS
Kubernetes in Depth