Agent Frameworks Research by Ry Walker
A comparison of 9 agent coordination protocols, layers, and frameworks — from the A2A standard (v1.0 under the Linux Foundation) and the layered protocol stack (AG-UI, Zed ACP, AGNTCY) to orchestration frameworks (CrewAI, AutoGen, LangGraph). Who coordinates agents, and how?
Compare 16 agent frameworks by execution model, state and recovery, approvals, language support, observability, licensing, and deployment responsibilities.
A comparison of 10 tools for agent memory and self-improvement — from memory layers like Mem0, Letta, Hindsight, and Cognee to evolving playbook systems like ACE, plus the research foundations driving the category.
A comparison of 15 agentic skills methods, catalogs, and registries, with current workflow changes, licensing and portability limits, and adjacent execution platforms.
Category analysis of 7 platforms that organize AI agents as employees in company structures — from Paperclip's 70k-star org chart orchestration to Edict's Tang Dynasty governance model. Covers the category's consolidation: who's growing, who's dormant, and who got reclassified.
Comparison of 19 autoresearch tools for program experiments, deep research, and scientific discovery: architecture, evaluation, access, costs, and limitations.