Securing Claude Agents: Sandboxing and Injection Defense
Harden Claude Managed Agents: sandbox tools, least-privilege scopes, keep secrets off the model, and contain prompt injection so hijacks stay harmless.
9 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 323 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
Cheaper, Faster Claude Agents: Caching and Batching
Reduce token cost and latency in Claude Managed Agents with prompt caching, batching over fan-out, lean context, and matching the model to each task.
9 min read
Debugging Claude Agents: Loops, Wrong Tools, Bad Args
Diagnose and fix the three failure modes in Claude Managed Agents — loops, wrong tools, and hallucinated arguments — with transcript-driven debugging.
10 min read
Prompt & Context Design for Claude Managed Agents
What to put in a Claude agent's context and what to leave out: the inclusion decision, retrieval over inclusion, durable state, lean handoffs.
8 min read
Wiring MCP Servers & Tools into Claude Agents
Connect MCP servers to Claude agents safely: server-side auth, tight schemas, typed errors, and idempotent writes that survive retries.
8 min read
Reusable Patterns for Claude Managed Agents
Code-level patterns for Claude agents: role contracts, guarded tools, context slices, structured results, and rubric-as-data verifiers.
8 min read
Build a Claude Managed Agent: Step-by-Step Walkthrough (Managed Agents Orchestration)
A hands-on walkthrough: build a Claude Managed Agent with subagents, tools, success criteria, budgets, and idempotent writes.
8 min read
Inside Claude Managed Agents: Architecture & Internals
How Claude Managed Agents work end to end: the orchestration loop, subagent isolation, durable state, verification, and budgets explained.
8 min read
Inside Claude Opus 4.7: The 1M Context Window Architecture Explained
A practical engineering deep dive into Claude Opus 4.7, covering architecture, tradeoffs, and what production teams need to know about 1M context window.
6 min read