| name | 4-agent-orchestration |
| description | 4-Agent orchestration model for AOS v2.0. Grok as Agent 1 (Coordinator) directs three specialized agents using the full framework and foundational skills. |
| version | 2.3 |
AOS v2.0 — 4-Agent Orchestration Model
Agent 1: Grok — Master Coordinator (Build Coordinator)
- Central router and decision maker.
- Always invokes foundational skills first (
xai-reflection, evidence-gate, mission-spine-guard, anti-pattern-scanner, paste-bridge-validator).
- Delegates to Agents 2–4 with clear handoff format.
- Performs final gates and synthesis.
Agent 2: Reasoning & Evidence Agent
- Specializes in deep analysis, research, and synthesis.
- Heavy use of
evidence-gate + xai-reflection + agnostic-evidence-analyst.
Agent 3: Verification & Integrity Agent
- Specializes in quality gates, anti-pattern detection, and Mission Spine enforcement.
- Heavy use of
anti-pattern-scanner, paste-bridge-validator, mission-spine-guard, mission-integrity-auditor.
Agent 4: Execution & Delivery Agent
- Specializes in building artifacts, writing, and delivery.
- Uses
working-doc-manager and execution-focused skills.
See full details in AGENT_PROMPTS.md (customization prompts for all 4 agents) and the main orchestration document.
This skill enables structured multi-agent collaboration with mandatory foundational skill usage and Mission Spine protection.
