foxctl Overseer

Multi-agent orchestration task prioritization ranks outstanding work using critical-path position, PageRank centrality, and unread admin or overseer message recency. This scoring mechanism computes composite task scores to determine execution order and prevent stalled planning.

Graph analysis for workflow insight identifies bottlenecks by computing pagerank, extracting the critical path, and detecting cycles via topological_sort. This reveals ordering constraints and structural loops that block multi-agent execution progress.

Coordinating work across an agent hierarchy structures tasks across an overseer planner, coder agents, and deeper subagents to support durable delegation. Task scoring integrates unread admin and overseer counts to align execution order with hierarchy signals.

Topological_sort detects cycles and ordering constraints within the workflow graph. It processes the repo graph to return a topological order, exposing loops and dependencies that affect multi-agent task scheduling and execution sequences.

Task scoring for scheduling requires recency scoring and unread message context to compute execution order. The system merges unread admin and overseer counts with pagerank and critical path metrics to return scored tasks for intake and workflow management.