Claude Agent Patterns for Security and Compliance Code
Reusable patterns for Claude security agents: typed tools, propose-gate-execute, idempotent writes, lean context, and tests that act as controls.
8 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 51 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
Claude Opus Security Agent Patterns You Can Reuse
Reusable code-level patterns for Claude Opus security agents: evidence ledgers, narrow tools, compaction subagents, context layering, and post-checks.
7 min read
Build a Claude Opus Security Triage Agent: Walkthrough
Step-by-step: build a Claude Opus security triage agent with the Agent SDK — tools, skills, gated actions, budgets, and your first live investigation.
8 min read
Connect Claude to a Security Tool: Step-by-Step Build
A follow-along build for wiring Claude to a security scanner over MCP: read-only first, then gated remediation with a full audit trail.
8 min read
Claude Security Tool Architecture: How the Pieces Fit
How Claude connects to security and compliance tools end to end: MCP boundaries, skills, policy gates, and an audit path that stays provable.
9 min read
How Claude Opus Powers a Security Agent: Architecture
How a Claude Opus cybersecurity agent fits together end to end — the model, MCP servers, skills, and a governed control layer for safe triage.
8 min read
Where Claude Code Single-File HTML Output Heads Next
Self-verifying artifacts, richer single-file tools, MCP-connected interfaces — where Claude Code's HTML pattern is heading and how to prepare your team.
8 min read
Where Claude Cowork Sales Books Are Heading Next
Where agentic sales books on Claude Cowork are heading — more autonomy, multi-agent orchestration, richer memory — and how to prepare now.
7 min read
How to Measure a Claude Cowork Sales Book Is Working
The metrics and signals that prove a 4,000-account Claude Cowork book is working — leading indicators, quality measures, and vanity-metric traps.
7 min read