heady-agent-runtime

Orchestrate distributed HeadyBee agents with phi-scaled quotas and CSL scheduling.

1|Updated Mar 24, 2026
One-click install
npx skills add https://github.com/HeadyAI/heady-context --skill heady-agent-runtime
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: heady-agent-runtime
Source: https://github.com/HeadyAI/heady-context/tree/main/heady-skills/heady-agent-runtime
Command: npx skills add https://github.com/HeadyAI/heady-context --skill heady-agent-runtime

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agent process runtime treating HeadyBee agents as first-class distributed workloads with phi-scaled resource quotas, CSL-scored preemptive scheduling, V8-isolate fault domains, and control-plane orchestration. Each bee receives CPU/memory/token budgets from Sacred Geometry pools (Hot 34%, Warm 21%, Cold 13%, Reserve 8%). Higher-coherence tasks preempt lower ones. Agents follow a six-state lifecycle (SPAWNING → READY → RUNNING → SUSPENDED → RETIRING → TERMINATED). Fault domains group agents by Sacred Geometry layer so failures never cascade cross-layer. The control plane tracks 89+ bee types, auto-decomposes complex tasks into sub-agent DAGs, and triggers semantic backpressure when quotas are exceeded.

Core Features & Use Cases

  • Six-state agent lifecycle (SPAWNING → READY → RUNNING → SUSPENDED → RETIRING → TERMINATED)
  • Phi-math based resource pools with per-bee budgets
  • CSL-scored preemptive scheduling prioritizing higher-coherence tasks
  • Fault-domain isolation across Sacred Geometry layers
  • Dynamic task decomposition into sub-agent DAGs and backpressure when quotas are reached

Quick Start

Spawn a new HeadyBee agent by sending a spawn request with the bee type and resource pool.

Frequently Asked Questions about heady-agent-runtime

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is preemptive scheduling for distributed agent fleets?

Fault-domain isolation groups distributed agents by Sacred Geometry layers so failures never cascade cross-layer. This V8-isolate based approach prevents faults in one layer from affecting agents in another layer.

How do I manage the lifecycle of a distributed agent from spawning to termination?

Resource allocation constraints trigger semantic backpressure when agents exceed their allocated phi-scaled CPU, memory, or token quotas. The system enforces these limits by suspending lower-coherence tasks to maintain pool stability.

Does this agent runtime support automatic decomposition of complex tasks into sub-agents?

Yes, the agent runtime supports automatic task decomposition. The control plane auto-decomposes complex tasks into sub-agent directed acyclic graphs (DAGs), distributing workloads across 89+ bee types within Sacred Geometry resource pools.

What's the best way to allocate CPU and memory quotas across distributed agent pools?

You scale distributed agent workloads by deploying them across Sacred Geometry pools with predefined phi-scaled quotas. The preemptive scheduler dynamically manages resource distribution, handling 89+ bee types and triggering backpressure when quotas are reached.