Agentic patterns

Twelve Agentic Design Patterns.

The pattern catalogue answers how agents coordinate. Agentic Design Patterns answer what failure mode an architecture stabilizes against — twelve patterns across four phases of one runtime loop.

Foundational phase · 3 patterns

World Modeling & State

Stabilizes the agent's grip on reality — every observation, memory access, and state transition entering the system is clean, salient, and durable. If that foundation is sand, no amount of clever planning or execution repairs it.

Validate → retrieve → persist

01
Integrator
AddressesCognitive data quality — the systemic weakness of trusting raw observations.
Validates incoming observations before they enter the world model, ensuring downstream reasoning operates on clean signals rather than raw, potentially malformed inputs.
When it stabilizes
When an agent must combine multiple noisy sensors, parse external APIs whose schemas drift, or admit user input that may misrepresent the world.
Example
A schema-validated JSON parser sits between a tool's raw response and the agent's state, rejecting malformed payloads before they poison the graph.
1related patternopen →
02
Retriever
AddressesInefficient retrieval — the failure mode of either flooding the prompt or missing the relevant fragment.
Context-sensitive interface to long-term memory; selects what is salient to the current step rather than dumping the entire store into the prompt.
When it stabilizes
When the knowledge base grows beyond a few pages and naive similarity search starts returning noise, or when different tasks need different retrieval strategies.
Example
A RAG retriever with a reranker selects top-k passages keyed to the current sub-goal, not the conversation history.
1related patternopen →
03
Recorder
AddressesPersistence — the loss of progress when runs span steps, sessions, or processes.
Saves and restores Reasoning & World Model (RWM) states for durability, replay, and forward recovery across steps, sessions, and processes.
When it stabilizes
When a workflow must survive restarts, when HITL pause points need durable state, or when concurrent users must not bleed into each other.
Example
Checkpoint snapshots of agent state at every node boundary, persisted to Postgres with a composite thread key of UserID + SessionID.
1related patternopen →

Explore the Foundational phase →

Cognitive phase · 3 patterns

Reasoning

Stabilizes the quality of step-by-step decision-making — where goals are prioritised, decomposed, and matched to actions with deliberate intent. A perfect executor running a flawed plan simply reaches the wrong destination faster.

Which goal → how → what action

01
Selector
AddressesGoal-selection under multiple competing demands.
Dynamic prioritization of goals — the tactical step-selector that picks the next action from competing demands.
When it stabilizes
When an agent receives simultaneous requests, when user intent is ambiguous, or when resource constraints force trade-offs.
Example
A priority queue ranks pending tasks by deadline and confidence, surfacing the highest-urgency item to the planner.
1related patternopen →
02
Planner
AddressesReasoning depth — the failure mode where one-shot reasoning collapses for tasks needing many steps.
Strategic decomposition of complex goals into ordered sub-goals that can be executed sequentially or in parallel.
When it stabilizes
When a task requires more than three sequential decisions, when failure at one step demands backtracking, or when parallel workstreams must be coordinated.
Example
A ReWoo-style plan-first module emits a DAG of sub-tasks before any tool is called, making the execution path inspectable.
2related patternsopen →
03
Deliberator
AddressesAction quality at each step.
Selection of the optimal action per planning step — the local choice given the local context, not a global default.
When it stabilizes
When multiple valid actions exist for a given state, when the cost of a wrong action is high, or when the context changes between steps.
Example
A Tree-of-Thoughts branch evaluates three candidate actions against a reward model before committing to the best one.
2related patternsopen →

Explore the Cognitive phase →

Execution phase · 3 patterns

Action

Stabilizes the bridge from reasoning to the world — every planned action carried out reliably, observed faithfully, and communicated safely. A perfect plan badly executed is indistinguishable from a bad plan — the user sees the same failure.

Run → bound tools → coordinate

01
Executor
AddressesAction reliability — handling failure, retry, and outcome capture.
Reliable execution and systematic feedback collection, with structured retry and outcome capture.
When it stabilizes
When external APIs are flaky, when partial failures must be distinguished from total failures, or when downstream nodes need the result to proceed.
Example
A LangGraph node wraps every tool call in a retry loop with exponential backoff, logging the exact response or error into shared state.
1related patternopen →
02
Tool Use
AddressesSafe tool boundaries — argument validation, capability scoping, error normalization.
Proxy / adapter interface for safe external function calls, enforcing argument validation and capability scoping.
When it stabilizes
When tools are provided by third parties, when argument schemas drift, or when a tool's failure mode could corrupt agent state.
Example
An MCP client adapter validates every argument against the server's JSONSchema before dispatch, returning a normalized error if validation fails.
2related patternsopen →
03
Coordinator
AddressesCommunication structure between agents.
Management of structured multi-agent communication, ensuring message flow is inspectable and bounded.
When it stabilizes
When more than two agents exchange messages, when message ordering matters, or when agents need to share partial results without exposing full state.
Example
A supervisor pattern routes messages through a central dispatcher that logs every handoff and enforces a max-handoff limit to prevent unbounded loops.
2related patternsopen →

Explore the Execution phase →

Adaptive phase · 3 patterns

Evolution

Stabilizes the agent against its own history — failure produces learning, experience compounds into skill, and guardrails don't erode over time. An agent that does not adapt costs more every run — it repeats mistakes and slowly loses its guardrails.

Diagnose → compound → monitor

01
Reflector
AddressesCausal learning — the gap between observing failure and understanding its root cause.
Causal failure analysis for strategy adjustment — the gap between "it failed" and "it failed because...".
When it stabilizes
When an agent repeatedly fails the same task, when success rates drop over time, or when the environment changes and old strategies break.
Example
A Reflexion node appends a causal critique to memory after every failed run, tagging the root cause so the planner can avoid the same decomposition next time.
2related patternsopen →
02
Skill Build
AddressesSkill compounding — the failure to convert one-off solutions into reusable assets.
Extraction of reusable procedures from past experience, converting one-off solutions into persistent assets.
When it stabilizes
When the same sub-task appears across multiple workflows, when prompt engineering is repeated for similar problems, or when agents must operate in low-latency regimes that preclude re-planning.
Example
A tool-calling pattern that succeeded ten times is extracted into a parameterized function template and registered in a skill library for future retrieval.
1related patternopen →
03
Controller
AddressesValue alignment and ongoing safety drift.
Continuous monitoring of ethical and operational guardrails, catching drift before it cascades.
When it stabilizes
When agents handle sensitive data, when outputs affect real-world decisions, or when guardrails defined at initialization are expected to hold across long-running sessions.
Example
An LLM-as-judge evaluator runs after every tool call, scoring the output against a safety rubric and halting the graph if the score drops below a threshold.
1related patternopen →

Explore the Adaptive phase →

Keep going

The patterns are the nouns. The loop is the grammar.

PatternsPattern OverviewAll four pattern categories on one map.All four categories PatternsThe Six Coordination PatternsThe six patterns every architecture is built from.Browse the catalog ReferenceGlossaryEvery term in one austere, scannable lookup.Open the glossary