Routes incoming requests to appropriate AI team members via Agent protocol.
npx skills add https://github.com/jedrecord/hal2026 --skill hal
Please help me install this Agent Skill. Skill: hal Source: https://github.com/jedrecord/hal2026/tree/main/skills/hal Command: npx skills add https://github.com/jedrecord/hal2026 --skill hal
Hal acts as the single point of contact for all requests, ensuring work is delegated to the correct AI team member and that results are tracked and reported instead of being executed directly.
Tell Hal to route a task by issuing the /hal command followed by your description.
High-intent search queries and answers about installing and using this skill.
AI agent orchestration routes incoming requests to the appropriate team member based on roster capabilities, dispatching work via the Agent protocol while tracking outcomes through the completion-status protocol.
Delegation guardrails prevent the orchestrator from executing tasks directly, ensuring work is strictly dispatched to specialized agents based on enforced routing rules and tracked project status.
Multi-step task delegation reads the agent roster and capabilities to dispatch work sequentially, using Learnings.md and project-status context to avoid conflicts and report final outcomes.
Context-aware orchestration uses Learnings.md and project-status files to track ongoing work, preventing delegation conflicts and ensuring proper reporting via the completion-status protocol.
Executing tasks directly violates enforced delegation guardrails; the orchestrator must read the roster, dispatch work via the Agent protocol, and report outcomes using the completion-status protocol.
Yes, agent workflow delegation requires a defined roster and agent capabilities file to correctly route requests, enforce guardrails, and track multi-step task outcomes.
OliverOuyangCatalog OMC agents, tools, pipeline stages, and git trailer conventions.
fcamacholMonitor registration process statuses and trigger change notifications.
nnexaiAutomate multi-phase project milestones through a discuss-plan-execute loop.
HomericIntelligenceDecompose complex tasks into hierarchical agent swarms with tiered model assignments.
