Inside T800: Contrarian Reviews Handled by a Filter Agent Before Dispatch
lucasmeijer · x · 2026-10-22
lucasmeijer asks filpizlo about the design of his T800 multi-agent system: does a "contrarian review" mean a different model family, or just a prompt that tells it to be a pain? How many lines does a typical review change? And does the filter agent get fresh context and re-read files from disk?
The quoted post outlines the architecture: the T800 top-level agent passes the contrarian review to a filter agent, which decides whether the review is good or goes too far, edits it if needed, and returns it to the top-level agent, which then dispatches work to work agents.
Related event: Developers debate multi-agent code review workflow designs(3 posts)→
More from coding & agent
- Greptile v5 spins up sub-agents on every PR, pairing open and frontier models — togethercompute · 2026-10-22
- Engineer turns his evals-to-business skill optimization loop into an AI-made Netflix-style documentary — kieranklaassen · 2026-10-22
- How to Build an AI Agent Harness from Scratch in TypeScript: freeCodeCamp's deep tutorial — adnan_hashmi · 2026-10-22
- Four cloud agents in two months: how Claude Duet, Grok Bot, Muse, and Dots actually differ — dotey · 2026-10-22
- The Ontology Pipeline: a semantic knowledge management framework for LLMs — adnan_hashmi · 2026-10-22
- Anthropic Duet preview sparks debate: is the "I built my own harness" crowd facing sunk cost? — altryne · 2026-10-22