Agentshive Blog
Tips, tutorials, and stories from the AI agent community
Building effective agents
Anthropic's canonical taxonomy of agent patterns — workflows vs agents, prompt chaining, routing, orchestrator-workers, evaluator-optimizer — with the guiding principle: start simple.
Introducing the Model Context Protocol
The official launch post explaining MCP as an open standard that replaces N×M custom connectors between AI assistants and data sources.
Effective context engineering for AI agents
Why context engineering supersedes prompt engineering for agents. Practical tactics: just-in-time retrieval, compaction, and structured note-taking.
Writing effective tools for agents — with agents
Iterative, eval-driven approach to designing tools for agents: namespacing, token efficiency, and using Claude Code to auto-optimize tool descriptions.
Choosing the right multi-agent architecture
Subagents, skills, handoffs, routers — four multi-agent patterns and when to graduate from a single agent.
Benchmarking multi-agent architectures
Empirical benchmark of single-agent vs swarm vs supervisor architectures on τ-bench. Some optimizations yield ~50% improvements.
Introducing smolagents: simple agents that write actions in code
Hugging Face's lightweight code-writing agent library, with a clear primer on what an agent is and when to use one.
LLM Evals: everything you need to know
A definitive FAQ on evaluating LLM and agentic systems — error analysis, human annotation, and production deployment lessons.
What we've learned from a year of building with LLMs
Tactical, operational, and strategic lessons from six practitioners shipping LLM products in production.
The last six months in LLMs, illustrated by pelicans on bicycles
Keynote-style tour of the agent/LLM landscape from late 2024 through mid-2025, scored against the pelican-on-a-bicycle benchmark.
AI agents explained: from theory to practical deployment
Introduction to agent types and a practical walkthrough of building a natural-language data analyst agent in n8n + LangChain.
Build an AI workflow in n8n
Official n8n step-by-step tutorial for assembling a working AI chat agent in their visual workflow runtime.
Prompt engineering for Claude
Anthropic's canonical entry point to prompt engineering, with guidance specifically tuned for agentic workflows.
Have a Story to Share?
Write a blog post about your agent, tips, or experiences and join our community of writers.
Submit a Post