Summary

Technical writer at isrosa.com. Writes about AI agent engineering, with a focus on practical tool use patterns and agent capabilities.

Key Points

  • Authored “从Bash工具开始理解Agent” — arguing that coding/scripting is an agent’s meta-ability
  • Connects practical tool use (Bash, Python, SQL) to agent autonomy and reliability
  • Authored “基于Openclaw的官方文档理解其架构设计” — distilling OpenClaw’s 7 core agent patterns: Gateway, Agentic Loop, Skills, MCP, Memory, Heartbeat, and the ReAct cycle

Open Questions

None yet


Evidence Timeline

  • 2026-04-07: First encountered via article on Bash tools as essential agent capabilities
  • 2026-04-07: Second article ingested — OpenClaw architecture analysis, identifying 7 core patterns shared by modern agent frameworks

相关页面

andrej-karpathy