Prompt template uses parameterized construction, not string concatenation

Scattered ad-hoc prompt construction across route handlers makes it structurally impossible to audit what user input reaches the model. When prompt building is decentralized, a single developer's shortcut—interpolating a query param into the system message in one handler—creates an injection vulnerability that is invisible to reviewers looking at other handlers. CWE-1427 and OWASP LLM01:2025 both identify inadequate prompt construction as a root-cause enabler. Centralized, typed prompt builders create a single auditable boundary: every change to what enters a prompt is visible, testable, and reviewable in one place. For teams under NIST AI RMF governance, demonstrating a controlled prompt construction process requires exactly this kind of centralization.

Create a dedicated prompt construction module (src/lib/prompts.ts) that accepts typed parameters and returns a fully-formed messages array. All AI call sites import from this module; no prompt building happens inline.

// src/lib/prompts.ts
import type { ChatCompletionMessageParam } from 'openai/resources'

const SYSTEM_PROMPT = `You are a helpful assistant for Acme. [full instructions]`

export function buildChatMessages(params: {
  userMessage: string
  context?: string // pre-sanitized retrieval context only
}): ChatCompletionMessageParam[] {
  return [
    { role: 'system', content: SYSTEM_PROMPT },
    ...(params.context
      ? [{ role: 'system' as const, content: `<context>\n${params.context}\n</context>` }]
      : []),
    { role: 'user', content: params.userMessage }
  ]
}

The TypeScript signature acts as a contract: adding a new variable input requires an explicit parameter, making it visible in code review rather than buried in a template literal.