Customerio Reference ArchitectureSAFE
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Overview
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
4f83675ca38aOBSERVED · 2026-10-08Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: customerio-reference-architecture description: 'Implement Customer.io enterprise reference architecture. Use when designing integration layers, event-driven architectures, or enterprise-grade Customer.io setups. Trigger: "customer.io architecture", "customer.io design", "customer.io enterprise", "customer.io integration pattern". ' allowed-tools: Read, Write, Edit, Bash(npm:*), Bash(npx:*), Glob, Grep version: 1.14.0 license: MIT author: Jeremy Longshore <[email protected]> tags: - saas - customer-io - architecture - enterprise compatibility: Designed for Claude Code --- # Customer.io Reference Architecture ## Prerequisites - A data-flow owner, approved workspace/environment boundaries, event contract, consent model, and incident/support route. - Repository/configuration review for any change to credentials, webhooks, message templates, or customer-data processing. ## Output - A documented architecture mapping event producers, identity/consent, Customer.io workspaces, observability, ownership, and failure handling. - Clear trust boundaries and reversible integration points for delivery, replay, and incident response. ## Examples Document a producer that validates and idempotently emits a versioned event to the development workspace, with a consent gate before campaign entry and redacted metrics by environment. Promote the same contract through staging before production, retain a dead-letter/replay owner, and keep recipient identifiers and tokens outside logs and diagrams. ## Overview Enterprise-grade reference architecture for Customer.io: a service layer separating Track and App API concerns, event-driven processing with message queues, repository pattern for user-to-CIO sync, webhook event bus, and infrastructure as code. ## Architecture Principles 1. **Two Clients, Two Concerns** — `TrackClient` for behavioral data in, `APIClient` for messages out 2. **Event-Driven** — Message queues decouple your app from Customer.io API availability 3. **Idempotent Operations** — All writes safely retryable via content hashing 4. **Service Layer** — Business logic never calls Customer.io SDK directly 5. **Observability** — Every operation emits timing and error metrics ## Architecture Diagram ``` ┌─────────────┐ ┌───────────────────┐ ┌──────────────┐ │ Application │───>│ MessagingService │───>│ Track API │ │ Routes │ │ (service layer) │ │ identify() │ └─────────────┘ │ │ │ track() │ │ - identify users │ └──────────────┘ │ - track events │ │ - send txn emails │ ┌──────────────┐ │ │───>│ App API │ └───────────────────┘ │ sendEmail() │ │ │ broadcast() │ │ └──────────────┘ v ┌───────────────────┐ │ Event Queue │ ┌──────────────┐ │ (Redis/Kafka) │───>│ DLQ │ │ for reliability │ │ (failures) │ └───────────────────┘ └──────────────┘ ┌─────────────┐ ┌───────────────────┐ ┌──────────────┐ │ Customer.io │───>│ Webhook Handler │───>│ BigQuery │ │ Webhooks │ │ HMAC verification │ │ (analytics) │ └─────────────┘ │ Event routing │ └──────────────┘ ``` ## Instructions ### Step 1: Core Service Layer ```typescript // services/messaging-service.ts import { EventEmitter } from "events"; import { TrackClient, APIClient, SendEmailRequest, RegionUS, RegionEU } from "customerio-node"; interface MessagingConfig { siteId: string; trackApiKey: string; appApiKey: string; region: "us" | "eu"; } export class MessagingService extends EventEmitter { private track: TrackClient; private app: APIClient; constructor(config: MessagingConfig) { super(); const region = config.region === "eu" ? RegionEU : RegionUS; this.track = new TrackClient(config.siteId, config.trackApiKey, { region }); this.app = new APIClient(config.appApiKey, { region }); } async identifyUser(userId: string, attrs: Record<string, any>): Promise<void> { const start = Date.now(); try { await this.track.identify(userId, { ...attrs, last_seen_at: Math.floor(Date.now() / 1000), }); this.emit("identify", { userId, latencyMs: Date.now() - start }); } catch (err) { this.emit("error", { operation: "identify", userId, err }); throw err; } } async trackEvent( userId: string, name: string, data?: Record<string, any> ): Promise<void> { const start = Date.now(); try { await this.track.track(userId, { name, data }); this.emit("track", { userId, name, latencyMs: Date.now() - start }); } catch (err) { this.emit("error", { operation: "track", userId, name, err }); throw err; } } async sendTransactional( to: string, templateId: string, data: Record<string, any>, identifiers?: { id?: string; email?: string } ): Promise<{ delivery_id: string }> { const start = Date.now(); try { const request = new SendEmailRequest({ to, transactional_message_id: templateId, message_data: data, identifiers, }); const result = await this.app.sendEmail(request); this.emit("transactional", { to, templateId, latencyMs: Date.now() - start }); return result; } catch (err) { this.emit("error", { operation: "transactional", to, templateId, err }); throw err; } } async triggerBroadcast( broadcastId: number, data: Record<string, any>, options: { segment?: { id: number }; emails?: string[]; ids?: string[] } ): Promise<void> { await this.app.triggerBroadcast(broadcastId, data, options); this.emit("broadcast", { broadcastId })
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (0)
No findings outside the package's declared scope.
Gates applied: no_behavioural_pass.
4f83675ca38afull audit observations/trust-audit/skill/jeremylongshore__customerio-reference-architecture.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | 4f83675ca38a | SAFE | B | 89 | first audit |
Questions
What does the Customerio Reference Architecture skill do?
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Is Customerio Reference Architecture safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Customerio Reference Architecture access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
Which assistants does Customerio Reference Architecture work with?
Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (4f83675ca38a), read on 2026-10-08. The repository is watched, and a new audit runs when it changes — this is the first audit.