04.7.4 · Skill in AI-native engineering
Agents
Systems where a model plans and takes multiple steps with tools.
An agent loops: it decides, calls tools, looks at results and continues. Agents can do open-ended tasks but are harder to predict and test. Clear tools, limits and human checkpoints keep them useful.
Resources
2- articleStart hereBuilding effective AI agents ↗Anthropic · 2024 · 12 min
When a simple workflow beats an agent, and the handful of patterns that cover most real systems.
- articleGo deeperEffective context engineering for AI agents ↗Anthropic · 2025 · 14 min
Choosing what goes into a model's limited context, and loading the rest only when needed.
Tools
none yetTools for this topic are being researched.