Zero Trust in Practice: Shipping a Safe Claude Refund Agent
End-to-end walkthrough of building a zero-trust Claude refund agent: least-privilege tools, scoped tokens, policy gates, adversarial evals, and audit logs.
8 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 37 of 631
Agentic AI & LLMs
Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Latest articles
Risk Management for Claude Agent Orchestration Systems
Failure scenarios, blast radius, and containment patterns for multi-agent Claude orchestration systems. A practical risk model for production.
7 min read
Risk Management for LLM-Driven Source-Code Security
Map the failure scenarios and blast radius of Claude-driven source-code security, and learn the containment patterns that keep an AI reviewer safe.
7 min read
Risk Management for Zero-Trust Claude Agents in Production
Failure scenarios, blast radius, and containment patterns for production Claude agents. A practical zero-trust risk management guide for engineers.
8 min read
Skills Your Team Needs for LLM Source-Code Security
What AppSec engineers, developers, and leaders must learn to make Claude-assisted source-code security work — and how hiring shifts around it.
8 min read
Skills Your Team Needs to Orchestrate Claude Agents
The hiring and learning shifts for building an agent orchestration system with Claude: new roles, the core skill mix, and a realistic team ramp.
7 min read
Zero Trust for AI Agents: The Skills Teams Must Hire For
The skills and hiring shifts teams need to build zero-trust Claude agents: identity, policy, red-teaming, and observability competencies explained.
8 min read
Scaling Claude Agent Orchestration Across an Organization
Scale Claude agent orchestration from one team to many without chaos: shared skills, registries, platform ownership, and standards that prevent sprawl.
7 min read
Scaling LLM Code Security Across Your Whole Org
Go from one team using Claude for secure coding to many without chaos: shared security skills, federated ownership, central governance, and honest coverage.
7 min read