Corpus Agentis
The field book to agent ecosystems
The field book to agent ecosystems
Design · Topologies

Agent topologies

Topology is the shape of an agent system. It shows where decisions are made, what holds control, where tools and sub-agents sit, and how information moves through the system. Topology is what turns a collection of capabilities into an operating design.

Hub-spoke agent topology diagram

Dense reasoning core radiating to satellite specialists. Orchestrator + specialists.

e.g. Salesforce Agentforce
Sequential agent topology diagram

Linked nodes in a horizontal pipeline; each output feeds the next.

e.g. Document/data pipelines
Recursive agent topology diagram

Reason-act-observe loop converging toward a goal.

e.g. Claude Code, Devin
Multi-agent colony agent topology diagram

Micro-clusters of peers with filament hand-offs.

e.g. AutoGen group chat
Memory-first agent topology diagram

Concentric memory rings with retrieval filaments.

e.g. RAG / research agents
Monitoring agent topology diagram

Dormant bell with trailing sensors; wakes on signal.

e.g. Observability / alerting agents
Peer-to-peer (A2A) agent topology diagram

Nodes connect by proximity; no centre.

e.g. A2A agent networks
Event-driven agent topology diagram

Sparse dormant field with active pulse epicentres.

e.g. Lindy, Zapier Agents
Self-Consistency (N-sample) agent topology diagram

N parallel solvers explore sampled reasoning paths; a majority-vote aggregator collapses them to one consistent answer.

e.g. MASS Self-Consistency reference topology
Fully-Connected Debate agent topology diagram

N agents in a fully-connected mesh iterate arguments and counter-arguments each round until consensus. Every agent sees every other's proposal per round.

e.g. MASS Debate topology · Constitutional AI critic panels
Sidecar (parasitic tap) agent topology diagram

A side-agent rides beside the main agent chain, tapping events without being on the primary path. Typical memory-harvesting pattern.

e.g. Mem0 / Zep sidecars · Letta background observers
Coaching / Meta-Supervisor agent topology diagram

A read-only meta-agent sits above the main agent, observing actions and providing advisory feedback. Main agent decides whether to act.

e.g. LangGraph human-in-the-loop patterns · Claude teach mode
Swarm Dispatch (spawn-collect-die) agent topology diagram

A coordinator spawns N ephemeral workers, collects outputs, terminates them. Workers are short-lived and non-communicating.

e.g. Fable-5 dispatch · OpenAI parallel tool-calls
Judicial Committee (quorum) agent topology diagram

N judge-agents independently evaluate a proposal; a verdict aggregator applies a quorum rule (majority / unanimity / weighted).

e.g. MASS quorum voting · policy-decision agents
Adversarial Review agent topology diagram

Producer emits, critic rebuts, arbiter (optional) resolves. Loops until convergence or timeout. Widely used in code-review + reasoning-verification.

e.g. Anthropic red-team pairs · code-review agents
Classify · Confirm · Dispatch (CCD) agent topology diagram

Scalarity's signature control topology. Brain classifies intent as a structured tool_call (the only LLM step). One dedicated confirmation turn gates the transition, human-in-the-loop or programmatic. Dispatch is 100% deterministic code with the safety gate embedded inside, no LLM in the write path. Capability grows by adding routes, not by branching code.

e.g. Scalarity Alfred · Nodal factory · CCD-authored agents
Map-Reduce agent topology diagram

Fan-out across data partitions in parallel; a reduce node aggregates the outputs. Built-in parallelism.

Router / Dispatch agent topology diagram

Classify intent and route to exactly one of N branches: not all. A cheap triage front-door, distinct from Hub & Spoke.

Hierarchical Decomposition agent topology diagram

A planner recursively spawns sub-planners and workers: open-ended task breakdown where scope is unknown upfront.

Blackboard agent topology diagram

A shared workspace agents read and write opportunistically, without direct coupling. No fixed control flow.

Workspace-protocol agent topology diagram

One agent reads stage-specific instructions from numbered folders; each stage produces plain-text output that seeds the next. Filesystem is the orchestrator; no multi-agent framework required. Human review gates sit between stages.

e.g. Model Workspace Protocol (MWP), Van Clief & McDermott 2026
SOP-encoded workspace agent topology diagram

Named role-agents (PM, architect, engineer) coordinate through a shared workspace of structured artefacts (PRD, design, tasks, code) written as files. Each downstream agent consumes the previous agent’s artefact from the filesystem.

e.g. MetaGPT, Hong et al. ICLR 2024
Phase-per-command folder pipeline agent topology diagram

A single-agent workflow driven by a folder tree (.specify/) plus a fixed phase sequence: constitution, specify, plan, tasks, implement. Each phase reads and writes markdown artefacts in its own directory; the filesystem carries state between commands.

e.g. GitHub Spec-Kit
Assembly-line role pipeline agent topology diagram

Named role-stages pass a growing artefact along a strict linear pipeline. Each stage owns one phase (analyst → architect → engineer → QA), reads what the previous stage produced, and adds its own contribution before handing off. The SOP is the topology.

e.g. MetaGPT, Hong et al. ICLR 2024
Role-anchored pipeline agent topology diagram

A pipeline whose stages are role-agents, not tasks. The same message flows through role-typed handlers (product owner, architect, coder), each transforming it. Distinct from assembly-line by making the role identity the coordination unit rather than the sequence.

e.g. MetaGPT, Hong et al. ICLR 2024
Autonomous multi-agent society agent topology diagram

Multiple autonomous agents with persistent memory, planning and reflection sharing an environment, forming and updating relationships without a coordinator. Emergent coordination through observation, memory retrieval and self-driven action.

e.g. Generative Agents, Park et al. UIST 2023
Directed message-exchange graph agent topology diagram

Agents coordinate through a directed graph of message exchanges. Any agent can address any other; the exchange history is the coordination substrate. Reads as a graph, not a chain, because messages fork, reply and cross-cut.

e.g. AutoGen, Wu et al. arXiv 2023
Manager-mediated group chat agent topology diagram

A manager (or moderator) agent runs a shared conversation of N specialist agents, selecting who speaks next based on the transcript. Coordination is centralised at the moderator; the transcript is the shared state.

e.g. AutoGen GroupChat, Wu et al. arXiv 2023
Two-agent role-play agent topology diagram

Two agents adopt paired roles (user + assistant, teacher + student, critic + generator) and iterate through structured dialogue. The dyad’s rules are the topology; a single task can traverse the loop many times.

e.g. CAMEL, Li et al. NeurIPS 2023
Deliberate branching search over states agent topology diagram

Reasoning states form a tree the agent expands and evaluates node-by-node, backtracking from dead ends. Search discipline (BFS or DFS) plus a heuristic evaluator select which branch to grow. Distinct from chain-of-thought by choosing between alternative continuations rather than committing to one.

e.g. Tree of Thoughts, Yao et al. NeurIPS 2023
Seeded by · peer-reviewed academic sources Live data view 30 sources
OFA-MAS, One-for-All Multi-Agent Topology Architecture Space
The Web Conference (WWW 2026) · ACM
T1 · canonical topologies
MASS, Multi-Agent System Search
Joint research · Google Research × University of Cambridge
T1 · topological pruning
MASEval, Framework-Agnostic Multi-Agent System Evaluation
ACL Demonstrations 2026 · aclanthology.org
T1 · declarative topology schema
REALM-Bench, Multi-Agent Planning Benchmark
ACM Digital Library
T1 · planning under uncertainty
ReAct
ICLR 2023
T1 · interleaved reasoning + tool actions
Reflexion
NeurIPS 2023
T1 · verbal self-reflection
Toolformer
NeurIPS 2023
T1 · self-supervised tool-use
MetaGPT
ICLR 2024
T1 · SOP-encoded roles
Voyager
arXiv (2023)
T1 · lifelong learning
Generative Agents
UIST 2023
T1 · believable agent society
AutoGen
arXiv (2023)
T1 · conversable agents
CAMEL
NeurIPS 2023
T1 · role-playing agents
Tree of Thoughts
NeurIPS 2023
T1 · deliberate search over reasoning paths
Attention Is All You Need
NeurIPS 2017
T1 · transformer architecture
Interpretable Context Methodology
arXiv 2603.16021 · Eduba / University of Edinburgh
T1 · Model Workspace Protocol (MWP)
Open Knowledge Format (OKF) v0.1
Google Cloud Blog · Data Analytics / BigQuery Tech Leads
T1 · portable knowledge-exchange standard
Outrageously Large Neural Networks
ICLR 2017 · Google Brain
T1 · sparse gating
Switch Transformer
JMLR 2022 · Google Brain
T1 · top-1 routing
Fast Inference from Transformers via Speculative Decoding
ICML 2023 · Google Research
T1 · draft-and-verify decoding
Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
NeurIPS 2022 · Google Research
T1 · chain-of-thought prompting
Self-Consistency Improves Chain of Thought Reasoning in Language Models
ICLR 2023 · Google Research
T1 · sample-and-vote
Least-to-Most Prompting Enables Complex Reasoning in Large Language Models
ICLR 2023 · Google Research
T1 · problem decomposition prompting
Improving Language Models by Retrieving from Trillions of Tokens
ICML 2022 · DeepMind
T1 · RETRO
Gemini 1.5
DeepMind Technical Report 2024
T1 · sparse MoE
Training Compute-Optimal Large Language Models
NeurIPS 2022 · DeepMind
T1 · Chinchilla scaling laws
TPU v4
ISCA 2023 · Google
T1 · TPU v4
Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them (BIG-Bench Hard)
ACL 2023 Findings · Google Research + Stanford
T1 · BIG-Bench Hard (BBH)
MemGPT
COLM 2024 · UC Berkeley Sky Lab
T1 · two-tier hierarchical memory
MetaGPT
ICLR 2024 · DeepWisdom
T1 · SOP-encoded multi-agent workflow
Defeating Prompt Injections by Design (CaMeL)
arXiv 2025 · Google DeepMind
T1 · capability-based dual-model design
Agent archetypes · 10 patterns Live data

The capability and control patterns that cross-cut every topology: what an agent can do, and how its control flow is governed. From internal research (unvalidated; sources cited in the originals).

Cross-cutting capabilities · 4
Control / reliability · 6
Field notes

The shape is the pattern: orbs are sub-agents or steps, lines are control flow

Topology is to an agent what topography is to terrain: it describes the shape. Which parts exist, how they are arranged, and the path information takes through them. Drawn out, a topology is nodes and edges, with a model runtime at the core, sub-agents and capabilities around it, and data moving between them. It is the anatomy reduced to its arrangement.

Show more

Across thirty named patterns the striking thing is how singular most of them are: one agent directing a set of actions, tools or sub-agents. Genuinely multi-agent designs, where several agents work in tandem across a process, are still the minority, and are the more likely direction of travel. Sidecar is a useful middle case, where a smaller agent rides beside the main chain and taps events to harvest memory or coach, without sitting on the primary path or altering the flow. Twenty-three of the thirty cite a Tier-1 source, the canonical shapes seeded by peer-reviewed multi-agent research: OFA-MAS (WWW 2026), MASS (Google Research and Cambridge), MASEval (ACL 2026) and REALM-Bench.

From the corpus, curated by Brandon Chaplin
Common questions
What is an agent topology?

A topology is the shape of a multi-agent system: which agent talks to which, and who decides what happens next. The usual shapes are one orchestrator with specialists under it, a fan-out to parallel workers, and a group that argues its way to one answer. The shape decides how the system fails as much as how it works.

What is a hub-and-spoke agent topology?

One orchestrator in the middle, specialist sub-agents around it. The orchestrator holds the plan and hands each sub-task to whichever agent is suited to it. It is the most common production shape because there is one place to look when something goes wrong.

What is map-reduce for agents?

Split one task across several agents running at the same time, then merge their answers into one. It suits work that divides cleanly, like summarising 100 documents or grading 50 answers. It breaks down when the sub-tasks need to know what the others found.

What is a debate or ensemble of agents?

Several agents answer the same question independently, then a judge picks or merges the answers. Disagreement is the point: separate attempts catch each other's mistakes before anything is committed. It costs several times the tokens of a single pass, so it is used on high-value work rather than at scale.

What is an agent swarm?

Agents that self-organise with no central controller. Each follows local rules and the coordination emerges from their interaction rather than from a plan someone wrote. It scales well in theory and is hard to debug in practice, which is why it is rare in production.

Do I need more than one agent?

Usually not. A single agent with good tools handles most tasks, and Anthropic recommends starting with the simplest thing that works and adding agents only when it stops working (Anthropic). Multiple agents earn their cost when sub-tasks can genuinely run in parallel or need separate context.

Next in the learning path