Amazon Quick proposes an agent primitives taxonomy
Understand this story
The Amazon Quick team proposed an agent primitives taxonomy that breaks down agent work using four questions—When, What, Who, and How—corresponding to three types of primitives: triggers, agents, and capabilities. Under the team's classification, triggers fall into four kinds—human-initiated, agent-delegated, scheduled, and event-driven—and carry both activation timing and task instructions; agents are defined as persistent, learnable execution entities composed of instructions, capabilities, and learned context. Capabilities cover tools, skills, and workflows, distributed along a spectrum from "fully agentic to fully deterministic," with skills loaded through progressive disclosure.
AI overview based on source reports · Updated 2 hr ago
Coverage timeline
Follow the coverage to explore different perspectives.
- AWS Machine Learning BlogTaxonomy of agentic primitives in Amazon Quick
Amazon Quick's team describes a taxonomy of agentic primitives derived from four questions—When, What, Who, and How—covering triggers (human-initiated, agent-delegated, scheduled, event-driven), agents as persistent learning actors with instructions, capabilities, and learned context, and capabilities spanning tools, skills, and workflows. The post argues skills and workflows retain value independent of model improvements, with skills able to crystallize into workflows and workflows dissolve back into skills, and notes a request like monitoring Slack for a daily digest is usually a trigger added to an existing agent. It is the first post in a series, with upcoming posts on the agentic operating system, background learning, and proactive feeds.
Activity for this story
Not enough continuous observations to show a trend yet.