← curriculum/phase 14 · Agent Engineering
edit on github ↗
← all lessons in phase
phase 14 · 42 lessons
  • 01The Agent Loop
  • 02Rewoo Plan And Execute
  • 03Reflexion Verbal Rl
  • 04Tree Of Thoughts Lats
  • 05Self Refine And Critic
  • 06Tool Use And Function Calling
  • 07Memory Virtual Context Memgpt
  • 08Memory Blocks Sleep Time Compute
  • 09Hybrid Memory Mem0
  • 10Skill Libraries Voyager
  • 11Planning Htn And Evolutionary
  • 12Anthropic Workflow Patterns
  • 13Langgraph Stateful Graphs
  • 14Autogen Actor Model
  • 15Crewai Role Based Crews
  • 16Openai Agents Sdk
  • 17Claude Agent Sdk
  • 18Agno And Mastra Runtimes
  • 19Benchmarks Swebench Gaia
  • 20Benchmarks Webarena Osworld
  • 21Computer Use Agents
  • 22Voice Agents Pipecat Livekit
  • 23Otel Genai Conventions
  • 24Agent Observability Platforms
  • 25Multi Agent Debate
  • 26Failure Modes Agentic
  • 27Prompt Injection Defense
  • 28Orchestration Patterns
  • 29Production Runtimes
  • 30Eval Driven Agent Development
  • 31Agent Workbench Why Models Fail
  • 32Minimal Agent Workbench
  • 33Instructions As Executable Constraints
  • 34Repo Memory And State
  • 35Initialization Scripts
  • 36Scope Contracts
  • 37Runtime Feedback Loops
  • 38Verification Gates
  • 39Reviewer Agent
  • 40Multi Session Handoff
  • 41Workbench For Real Repos
  • 42Agent Workbench Capstone
‹
›
phase 14 · lesson 20 of 42

Benchmarks Webarena Osworld

20-benchmarks-webarena-osworld·full lesson folder ↗

← previous
Benchmarks Swebench Gaia
complete & next →
Computer Use Agents
Edit with