The 2025 World's Fair cemented 'agent engineer' as a job title — harness reliability, context engineering and evals dominated hallway track and stage alike.
The Best of Agent Harnesses list ranks 167 agent harnesses weekly by capability, autonomy and recovery, ships an MCP server so agents can pick their own harness, and maintains a graveyard that flagged a 268k-star repo for suspected star manipulation. Curation itself becomes harness infrastructure.
Anthropic makes Claude Code generally available and holds its first developer conference — with Claude Code for web, enterprise features, and hooks/MCP improvements central to agent workflows.
Anthropic open-sources the SDK behind Claude Code — the agent loop, tool use, subagents, memory and hooks — letting developers build their own Claude-powered agents on production rails.
OpenAI releases an open-source Agents SDK (successor to the Swarm experiments) with handoffs, guardrails and tracing, plus a Responses API purpose-built for agentic workflows.