How Temporal used one spec and agents to find hundreds of SDK integration defects
francesc · x · 2026-11-19
At PortCon, Temporal's Francesc Campoy Flores described scaling integration testing with agents. Temporal ships durable-execution plugins for OpenAI Agents, Google ADK, LangGraph, Strands and Pydantic AI. They wrote a single precise spec covering shared questions, letting coding agents generate runnable reproductions per plugin (many injecting deliberate failures). Basics held everywhere; agents found hundreds of defects, mostly in specific setups. Fixes landed within weeks, and lessons from one plugin became checks for the rest. Pitfalls included rules agents couldn't see and a review loop burning absurd tokens on one PR.
More from coding & agent
- Dev uses Grok bot as personal enforcer: Amazon refunds, parking fights and a VA headstone quest — doooyle · 2026-11-19
- Mnemos Ledger: Anchoring Your Agent's Memory On-Chain So It Outlives Its Machine — RileyRalmuto · 2026-11-19
- Vishal Sikka's Hang Ten raises $53M as AI deploycos cut delivery cycles from 90 to 14 days — vsikka · 2026-11-19
- Who pays when an AI agent buys the wrong thing? VGS exec: likely the merchants — schwentker · 2026-11-19
- Backprop Summit: how many Black Friday purchases will AI agents actually complete? — schwentker · 2026-11-19
- Databricks Launches Unity Gateway Insights to Auto-Find Coding Agent Cost Savings — matei_zaharia · 2026-11-19