Security Hardening for Claude Agents With Tool Access
Sandboxing, least privilege, secrets handling, and prompt-injection defense for Claude agents connected to security and compliance tools.
6 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 50 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
Cutting Claude Opus Token Cost for Security Agents at Scale
Prompt caching, batching, and model routing that keep Claude Opus security agents fast and cheap at thousands of runs a day.
8 min read
Cutting Token Cost on Claude Compliance Agents
Caching, batching, and result shaping that keep Claude security and compliance agents fast and cheap without sacrificing audit accuracy.
6 min read
Debugging Claude Agents Hooked to Security Tools
Fix loops, wrong tool calls, and hallucinated arguments when Claude agents connect to SIEM, scanners, and compliance tools — a practical debugging playbook.
7 min read
Debugging Claude Opus Security Agents: Loops & Bad Tool Calls
Fix the top Claude Opus agent failures — loops, wrong tool calls, and hallucinated arguments — with concrete debugging tactics for security workflows.
8 min read
Prompt and Context Design for Claude Opus Security Agents
Design a Claude Opus security agent's context: what to include, what to leave out, layering, compaction, and defending against prompt injection.
7 min read
Prompt and Context Design for Claude Security Agents
What to put in a Claude security agent's context, what to leave out, and why: lean, intentional, eval-tuned context design for compliance work.
8 min read
Wiring MCP Servers to Claude: Auth, Schemas, Errors
Production MCP engineering for Claude security tools: scoped auth, typed schemas, structured error handling, and idempotency that survive retries.
8 min read
Wiring MCP Tools into a Claude Opus Security Agent
Wire MCP servers into a Claude Opus security agent the right way: least-privilege auth, strict schemas, typed errors, and idempotent write actions.
7 min read