VSM Loop: Viable-System Orchestration for Multi-Agent R&D
The VSM Loop is a meta-workflow designed for managing hierarchies of subagents in research and development. It emphasizes the importance of planning and coordination among multiple decision points to ensure effective execution of large tasks. By implementing the Viable System Model, the loop aims to reduce rework and enhance the viability of systems at every level.
- ▪The VSM Loop orchestrates multi-agent R&D using Stafford Beer's Viable System Model.
- ▪It is particularly useful for large projects that require recursive sub-tasks and explicit decision-making processes.
- ▪The model includes five subsystems that work together to maintain the viability of the system.
Opening excerpt (first ~120 words) tap to expand
VSM Loop: Viable-System Orchestration for Multi-Agent R&D A meta-workflow for running hierarchies of subagents under Stafford Beer's Viable System Model (VSM). Use this when the work is large enough to span multiple research and build cycles, recursive sub-tasks, or long-running R&D where explicit discipline about who decides what, what crosses each channel in what form, and when to escalate to the user will pay back the overhead. This skill is the strategic and managerial layer. Inside it, compose [[review-loop]] for build–fix–review cycles and [[research-loop]] for investigation cycles. Sub-tasks substantial enough to be viable systems of their own follow the [[vsm-node]] recursion contract, so the hierarchy stays viable at every level.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at GitHub.