Agentic Mesh
Reference skill for the Agentic Mesh book (Falconer, O'Reilly 2025). Provides the theoretical and architectural foundation for building agent ecosystems.
Core Definitions
- Agentic Mesh: Interconnected ecosystem for agent discovery, collaboration, interaction, and transactions. Inherits from service mesh (APIs) and data mesh (data products).
- Agent: LLM-powered program that independently makes decisions, plans iteratively, and executes complex tasks. Distinguished from workflows by autonomous reasoning and nondeterministic execution.
- Agentic Quantum: Smallest meaningful unit = LLM (brain) + tools (limbs) + execution framework, containerized as a microagent.
Architecture at a Glance
Agent Anatomy (4 components)
| Component | Role |
|---|---|
| Brain (LLM) | Reasoning, planning, language — stateless, multimodal |
| Memory | Native / short-term / long-term / episodic / procedural / semantic |
| Context Engineering | Hot/warm/cold cache tiers, RAG, compression, slotting |
| Tools | Sensors + actuators, MCP protocol, tool chaining |
Agent Types (4)
| Type | Behavior |
|---|---|
| Task-oriented | Clear objective, execute plan, return result |
| Goal-oriented | Collaborative problem-solving, dynamic planning, shared workspace |
| Simulation | Virtual models, emergent behavior analysis |
| Observer | Continuous monitoring, event-driven, pub/sub |
Mesh Platform (6 components)
Registry, Monitor, Interactions Server, Marketplace, Workbenches (Consumer/Creator/Trust/Operator), Proxy
Agent Lifecycle
Draft -> Registered -> Published -> Certified -> Updated -> Deprecated -> Retired
Key Frameworks
Seven-Layer Trust Framework
| Layer | Domain |
|---|---|
| L1 | Identity & Authentication (cryptographic, mTLS) |
| L2 | Authorization & Access Control (RBAC/ABAC, OAuth2) |
| L3 | Purpose & Policies (data contracts, constraints) |
| L4 | Explainability (task plans, tool selection logic) |
| L5 | Observability & Traceability (logs, correlation, monitoring) |
| L6 | Certification & Compliance (evaluation, stress testing) |
| L7 | Governance & Lifecycle (bodies, ownership, escalation) |
Reliability Solution
- Problem: Combinatorial explosion — 0.99^1000 = 0.004% accuracy
- Solution: Task decomposition -> task independence -> specialization -> deterministic execution
Patterns Catalog
- Communication (6): Interaction, Delegation, Conversation, Attention, Broadcast, Listener
- Role (6): Planner, Orchestrator, Executor, Observer, Judge, Enforcer
- Organizational (8): Singleton, Team, Organization, Swarm, Ecosystem, Legal Entity, Federation, Supply Chain
- Messaging (7): Request-response, Async, Event-driven, Message queue, Streaming, Actor model, Shared workspaces
Operating Model (5 Pillars)
Structure (people + agents) | Process (lifecycle + governance) | Technology (registry, observability) | Policy (autonomy tiers, guardrails) | Metrics (value + safety)
Scaling Units
Agent (person) -> Fleet (team) -> Ecosystem (organization) -> Supply chain (economy)
Implementation Roadmap (5 Workstreams)
- Strategic foundations (vision -> MVP)
- Technology (build -> industrialize -> secure -> model ops)
- Agent & fleet factories (frameworks -> DevSecOps -> factories)
- Organization (operating model -> change management -> training)
- Governance (agent governance -> fleet governance)
Reference Files
Detailed concept files — read the specific file when deeper context is needed:
| File | Covers | When to read |
|---|---|---|
| foundations.md | Core definitions, mesh vision, enterprise requirements, scale predictions | Defining what agentic mesh is, why it matters |
| agents-landscape.md | AI history, workflows vs agents, evolution stages | Comparing agents to workflows, understanding agent evolution |
| agent-basics.md | Planning, execution, tools, memory, learning, collaboration | Understanding how individual agents work |
| agent-architecture.md | Principles, components, types, all pattern catalogs, state management | Designing agents, choosing patterns |
| enterprise-agents.md | Microagents, reliability, explainability, discovery, AgentOps, testing | Building production-grade agents |
| ecosystem-ux.md | Fleets, mesh components, lifecycle, marketplace, workbenches | Designing the mesh platform and UX |
| registry-interaction.md | Registry data model, conversations, interactions, events, super-contexts | Building registry and communication infrastructure |
| security-governance.md | Seven-layer trust, zero-trust, prompt injection, certification | Implementing security and governance |
| operations-factory.md | Operating model, team roles, fleet structures, agent factory, SDLC | Setting up operations and agent factories |
| practical-roadmap.md | Five workstreams, technology stack, dependencies, scale progression | Planning implementation |
Original Chapters
Full book source material lives in references/chapters/. The concept files above are curated summaries — go to chapters for verbatim detail.
| Chapter | File |
|---|---|
| Foreword | references/chapters/00-foreword.md |
| Preface | references/chapters/01-preface.md |
| Part I intro | references/chapters/02-part-i-defining-the-essentials.md |
| Ch01: Understanding Agentic Mesh | references/chapters/03-ch01-understanding-agentic-mesh.md |
| Ch02: Agentic Past, Present & Future | references/chapters/04-ch02-agentic-past-present-future.md |
| Ch03: Agents vs AI Workflow | references/chapters/05-ch03-agents-versus-ai-workflow.md |
| Ch04: Agent Basics | references/chapters/06-ch04-agent-basics.md |
| Part II intro | references/chapters/07-part-ii-defining-agent-ecosystem.md |
| Ch05: Agent Architecture | references/chapters/08-ch05-agent-architecture.md |
| Ch06: Enterprise-Grade Agents | references/chapters/09-ch06-enterprise-grade-agents.md |
| Ch07: Agentic Mesh Ecosystem | references/chapters/10-ch07-agentic-mesh-ecosystem.md |
| Ch08: Agentic Mesh UX | references/chapters/11-ch08-agentic-mesh-ux.md |
| Ch09: Agentic Mesh Registry | references/chapters/12-ch09-agentic-mesh-registry.md |
| Ch10: Interaction Management | references/chapters/13-ch10-interaction-management.md |
| Ch11: Security Considerations | references/chapters/14-ch11-security-considerations.md |
| Ch12: Trust Framework & Governance | references/chapters/15-ch12-trust-framework-governance.md |
| Part III intro | references/chapters/16-part-iii-building-your-agentic-mesh.md |
| Ch13: Operating Model & Team Structure | references/chapters/17-ch13-operating-model-team-structure.md |
| Ch14: Agent Factory | references/chapters/18-ch14-agent-factory.md |
| Ch15: Practical Roadmap | references/chapters/19-ch15-practical-roadmap.md |