Back to browse
GitHub Repository

your favorite harness deserves a chassis 🏎️ - multi-agent orchestration layer for experimental design

10 starsPython

Multi-agent orchestration layer for experimentation

by theo-kirbyΒ·May 17, 2026Β·1 pointΒ·0 comments

AI Analysis

●MidShip It

File-system driven agent config is clean, but OpenClaw and LangGraph already solve this.

Strengths
  • β€’Two-user filesystem isolation keeps secrets out of the agent's address space.
  • β€’Per-branch namespacing allows multi-tenant chassis operation on a single host.
  • β€’Pluggable LLM runtime supports swapping between Pi, Claude Code, or custom agents.
Weaknesses
  • β€’Zero stars and forks suggest an embryonic project lacking community validation.
  • β€’Reinvents orchestration wheels already turned by mature frameworks like LangGraph.
Category
Target Audience

AI researchers and developers building multi-agent systems

Similar To

LangGraph Β· OpenClaw Β· AutoGen

Post Description

chassis is a bare-bones agent orchestration layer designed around a two user agent-native file system running in docker. The goal is to streamline the agent/task hierarchy into its simplest form for maximum extensibility and minimal bloat.

Similar Projects