orchestration-framework

name	orchestration-framework
description	Core orchestration framework for model-agnostic multi-agent workflows with handoff protocol, policy governance, and configuration schemas
category	Orchestration Framework
metadata	{"version":"1.0","source":"Agent33","tags":["orchestration","multi-agent","handoff","policy","governance"]}

Orchestration Framework

This index connects the core orchestration system, AEP workflow, and reusable governance assets. Migrated from the Agent33 orchestration protocol and adapted for EVOKORE-MCP.

Model-Agnostic Principle

All guidance is written to be model-neutral. If a task requires a specific tool or model, document it in TASKS.

Sub-Skills

Skill	Description
handoff-protocol/	Structured handoff protocol for multi-session orchestration
policy-pack-v1/	Agent behavior policies, evidence requirements, and modular rules
schemas/	JSON schemas for agents, workflows, and orchestrator configuration

Orchestration (Handoff Protocol)

• handoff-protocol/STATUS.md - Runtime assumptions and constraints
• handoff-protocol/PLAN.md - Objectives and success criteria
• handoff-protocol/TASKS.md - Task queue and status
• handoff-protocol/DEFINITION_OF_DONE.md - Task completion checklist
• handoff-protocol/REVIEW_CAPTURE.md - Review capture template
• handoff-protocol/REVIEW_CHECKLIST.md - Review checklist with risk triggers
• handoff-protocol/SESSION_WRAP.md - End-of-session wrap template
• handoff-protocol/PRIORITIES.md - Rolling horizon priorities
• handoff-protocol/SPEC_FIRST_CHECKLIST.md - Pre-implementation scope checklist
• handoff-protocol/AUTONOMY_BUDGET.md - Per-task autonomy boundaries
• handoff-protocol/HARNESS_INITIALIZER.md - Session initialization protocol
• handoff-protocol/PROGRESS_LOG_FORMAT.md - Progress log format and rotation
• handoff-protocol/EVIDENCE_CAPTURE.md - Evidence capture template
• handoff-protocol/DECISIONS.md - Architecture decision log
• handoff-protocol/ESCALATION_PATHS.md - Escalation decision guide

Policy Pack

• policy-pack-v1/AGENTS.md - Agent behavior principles
• policy-pack-v1/ORCHESTRATION.md - Handoff protocol standard
• policy-pack-v1/EVIDENCE.md - Evidence requirements
• policy-pack-v1/RISK_TRIGGERS.md - Review trigger conditions
• policy-pack-v1/ACCEPTANCE_CHECKS.md - Acceptance checks per change type
• policy-pack-v1/PROMOTION_GUIDE.md - Asset promotion rules

Modular Rules

• policy-pack-v1/rules/README.md - Rules index and customization guide
• policy-pack-v1/rules/security.md - Security rules
• policy-pack-v1/rules/testing.md - Testing rules
• policy-pack-v1/rules/git-workflow.md - Git workflow rules
• policy-pack-v1/rules/coding-style.md - Coding style rules
• policy-pack-v1/rules/agents.md - Agent delegation rules
• policy-pack-v1/rules/patterns.md - Common patterns rules
• policy-pack-v1/rules/performance.md - Performance and efficiency rules

Configuration Schemas

• schemas/agent.schema.json - Agent definition schema
• schemas/workflow.schema.json - Workflow definition schema
• schemas/orchestrator.schema.json - Orchestrator configuration schema

AEP Workflow

The Align-Execute-Prove (AEP) workflow is the core operating model:

1. Align: Define scope and acceptance checks using SPEC_FIRST_CHECKLIST.md
2. Execute: Implement minimal changes, log progress using PROGRESS_LOG_FORMAT.md
3. Prove: Run verification, capture evidence using EVIDENCE_CAPTURE.md

Quick Start

1. Read handoff-protocol/STATUS.md + handoff-protocol/PLAN.md + handoff-protocol/TASKS.md to load context
2. Review policy-pack-v1/RISK_TRIGGERS.md to understand review requirements
3. Pick a task from handoff-protocol/TASKS.md and follow the task template
4. Use spec-first workflow: complete handoff-protocol/SPEC_FIRST_CHECKLIST.md
5. Capture evidence: use handoff-protocol/EVIDENCE_CAPTURE.md
6. Wrap session: complete handoff-protocol/SESSION_WRAP.md