Agent Design
4 articles about agent design.
You Don't Have a Claude Code Problem, You Have an Architecture Problem
·2 min read
When AI agents struggle with desktop automation, the issue is usually architecture - not the LLM. Thin action primitives that the model composes into workflows scale far better than monolithic scripts.
architectureclaude-codedesktop-automationprimitivesagent-designworkflows
Multi-Agent Hype vs Economic Reality in Production
·2 min read
A planner-executor-reviewer agent chain sounds elegant but burns 3x the tokens of a single well-prompted agent. Here is when multi-agent is worth it and when it is not.
multi-agenttoken-costsproductionai-economicsagent-designllm-costs
Octopus Cognition - Why AI Agents Split Brain from Arms
·2 min read
The best AI agent architecture separates reasoning (LLM) from execution (MCP tools) like an octopus. Each arm does its own perception through accessibility tree traversal.
ai-architecturemcpdistributed-cognitionagent-designmacos
How to Build AI Agents You Can Actually Trust - Bounded Tools and Approval UX
·3 min read
Giving AI agents broad system access is a recipe for disaster. How bounded tool interfaces and smart approval flows make desktop agents safe to use.
ai-safetyagent-designtrustuxdesktop-agent
Browse by Topic
Claude Code (101)Automation (94)Macos (79)Productivity (76)Ai Agent (74)Ai Agents (61)Desktop Agent (54)Parallel Agents (49)Accessibility Api (39)Tutorial (37)Developer Tools (34)Claude Md (31)Comparison (31)Mcp (29)Developer Workflow (27)Desktop Automation (26)Open Source (25)Memory (24)Privacy (22)Workflow (22)