Examples

AI Replenishment (Claude)

What this example shows

An AI actor using @loop-engine/adapter-anthropic analyzes inventory evidence and submits a replenishment recommendation. A confidence threshold gate keeps low-confidence decisions from reaching execution. A human approval transition remains required for final ordering.

Loop diagram

1IDLE ----[signal_detected]----> ANALYZING
2ANALYZING --[submit_recommendation]--> PENDING_APPROVAL
3ANALYZING --[insufficient_confidence]--> IDLE
4PENDING_APPROVAL --[approve]--> ORDERED (terminal)
5PENDING_APPROVAL --[reject]--> IDLE

Actors

| Actor | Type | Transitions | Guards | |---|---|---|---| | signal detector | automation | signal_detected | none | | claude recommender | ai-agent | submit_recommendation | confidence-threshold | | buyer approver | human | approve, reject | human-only |

Key annotated snippet

1import Anthropic from "@anthropic-ai/sdk";
2import { createAnthropicActorAdapter } from "@loop-engine/adapter-anthropic";
3 
4const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });
5const claude = createAnthropicActorAdapter(anthropic, {
6  modelId: "claude-opus-4-6",
7  confidenceThreshold: 0.75
8});
9 
10const { actor, decision } = await claude.createSubmission({
11  loopId: "scm.replenishment",
12  loopName: "SCM Replenishment",
13  currentState: "ANALYZING",
14  availableSignals: [{ signalId: "submit_recommendation", name: "Submit Recommendation" }],
15  instruction: "Recommend reorder quantity based on stock and forecast.",
16  evidence: {
17    sku: "LMB-BRS-001",
18    currentStock: 142,
19    reorderPoint: 280,
20    forecastedDemand: 527
21  }
22});
23 
24await engine.transition({
25  aggregateId: "repl-lmb-001" as never,
26  transitionId: "submit_recommendation" as never,
27  actor,
28  evidence: decision
29});

What emitted events look like

1{
2  type: "loop.transition.executed",
3  loopId: "scm.replenishment",
4  aggregateId: "repl-lmb-001",
5  transitionId: "submit_recommendation",
6  fromState: "ANALYZING",
7  toState: "PENDING_APPROVAL",
8  actor: { type: "ai-agent", id: "agent:claude-replenishment" },
9  occurredAt: "2026-03-13T12:24:00.422Z"
10}

Try it yourself

1cd loop-examples/ai-actors/claude
2pnpm install
3pnpm dev

See AI Replenishment (OpenAI) for the identical loop with a different provider adapter.