How To Tame Your Agents?
summary
Chronological account of building agentic pipelines from copy‑paste prompts to a multiagent factory, claiming a 16‑agent system built a C compiler and showing a working review agent demo that creates and reviews GitLab MRs using OpenCode, MCP, skills, sub‑agents, Playwright and Modbus; demonstrates MCP vs skills tradeoffs (determinism and context bloat vs light CLI skills) and the need for guardrails after spending ~₹5,000 on hallucinating runs; offers concrete patterns—plan/PRD, specialist agents, sub‑agents for builds—and shows the pipeline can work but costs and access controls matter.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:53Copy-paste factory approach
- 8:05Open source agent tools
- 12:00Skills vs MCP servers
- 15:55Live demo: review agent
- 21:17Plan, build, test, deploy
- 26:57Stripe’s factory insights
- 29:49Zero-touch agent vision
- 33:45Q&A
related talks
27mAgentic Coding with Discipline & Skill: Practical Techniques and Tips
01h 03mAI Coding Agents and how to code them
40mBuilding Agents with Agent Bricks and MCP
45mAI agents in practice - beyond the prototype
22mHOW TO BUILD YOUR OWN AI AGENT: THE COMPLETE ARCHITECTURE FOR ENGINEERING TEAMS
40mAgents in Action: Advancing Open Source AI Missions
01h 28mAI Coding Agents and how to code them
11mBuilding Secure Backend Services with AI Agents