Core Concepts

Agents and RAG

What Agent Frameworks Do

Agent frameworks like LangGraph, OpenAI Assistants, and custom agent loops focus on retrieving context, selecting tools, chaining operations, and generating recommendations from retrieved information.

They are excellent at reasoning. They are not designed to govern whether the resulting action is authorized, attributable, or auditable.

The Governance Gap

A typical agentic flow:

1Knowledge Base
2     ↓
3RAG Retrieval           <- context retrieval
4     ↓
5LLM / Agent Reasoning   <- recommendation
6     ↓
7Tool / API Execution    <- action

This works for low-risk autonomous tasks.

In enterprise production - supply chain, financial approvals, healthcare, infrastructure - the gap between "agent recommends" and "system executes" needs an explicit governance layer.

Where Loop Engine Fits

1Knowledge Base
2     ↓
3RAG Retrieval
4     ↓
5LLM / Agent Reasoning
6     ↓
7Loop Engine             <- decision governance
8     ↓
9Workflow Execution

The agent's recommendation becomes a proposed transition in a Loop Engine loop. Guards evaluate it. A human approves when policy requires. The action executes only after the loop allows it.

RAG Outputs as Evidence

RAG systems already produce high-value evidence: retrieved documents, confidence scores, and graph relationships.

Loop Engine attaches this context directly to transitions as structured evidence, creating a permanent record of what the model used, what it recommended, and why the system allowed the action.

Code Example

1"cmt">// @no-typecheck
2import Anthropic from "@anthropic-ai/sdk";
3import { createAnthropicActorAdapter } from "@loop-engine/adapter-anthropic";
4import { createLoopSystem } from "@loop-engine/sdk";
5 
6const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });
7const adapter = createAnthropicActorAdapter(anthropic, { modelId: "claude-opus-4-6" });
8const { engine } = await createLoopSystem({ loops: [/* procurement loop */] });
9 
10const ragResults = {
11  retrievedDocs: 12,
12  confidence: 0.84,
13  ragSource: "vendor-policy-kb"
14};
15 
16const { actor, decision } = await adapter.createSubmission({
17  loopId: "procurement",
18  loopName: "SCM Procurement",
19  currentState: "pending_analysis",
20  availableSignals: [{ signalId: "submit_recommendation", name: "Submit Recommendation" }],
21  instruction: "Recommend whether to proceed with purchase order issuance.",
22  evidence: { ...ragResults }
23});
24 
25await engine.transition({
26  aggregateId: "PO-10042" as never,
27  transitionId: "submit_recommendation" as never,
28  actor,
29  evidence: { ...decision, ragResults }
30});

Commerce Chain and Loop Engine

Loop Engine is the runtime that powers the Commerce Chain platform. Each @betterdata/scm-* and module declares how it participates in loops using a from —a typed map of which domain events the module handles and which loop IDs it joins.