Skip to content
Back to skills

Ramp Reference Architecture

ASecurity

Design a production Ramp integration with isolated OAuth clients, typed API adapters, durable webhook ingestion, reconciliation workers, and audited financial writes. Use when reviewing architecture or building a new platform. Trigger with "Ramp architecture" or "design Ramp integration".

  • 2,719 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 21, 2026
ai-agentsrustgoapisecuritydocumentation

Works with

  • cli
  • api
  • mcp

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned September 21, 2026

npx -y skills add jeremylongshore/claude-code-plugins-plus --skill ramp-reference-architecture --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Ramp Reference Architecture?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Ramp Reference Architecture
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/jeremylongshore-ramp-reference-architecture-7ec839b0/badge)](https://www.skillsdirectory.com/skills/jeremylongshore-ramp-reference-architecture-7ec839b0)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: ramp-reference-architecture
description: >-
  Design a production Ramp integration with isolated OAuth clients, typed API adapters, durable webhook ingestion, reconciliation workers, and audited financial writes. Use when reviewing architecture or building a new platform. Trigger with "Ramp architecture" or "design Ramp integration".
allowed-tools: Read,Glob,Grep,Write,Edit
version: 2.0.0
argument-hint: "[integration-or-environment]"
model: inherit
effort: high
author: Jeremy Longshore <jeremy@intentsolutions.io>
license: MIT
compatibility: Requires current Ramp Developer API documentation and approved access for any live financial, card, identity, accounting, application, or configuration change
tags: [saas, ramp, architecture, integration, governance]
---
# Ramp Integration Reference Architecture

## Overview

Separate trust and failure domains: credential broker, typed adapter, event ingress, durable work ledger, financial connectors, reconciliation, and evidence. Keep write authority narrower than read and observability paths.

## Prerequisites

- Identify the Ramp application, environment, business entities, affected data and workflows, accountable owner, and rollback boundary.
- Read `references/official-docs.md` and re-check endpoint schemas, scopes, limits, and support status before a live operation.
- Use synthetic fixtures or Ramp sandbox until production access and business effects are explicitly approved.
- Prepare approved secret storage and a sanitized evidence location.

## Current Contract

- Ramp supports direct Developer API, webhooks, MCP, and CLI surfaces for different automation jobs; long-lived deterministic integrations belong on the Developer API.
- Webhook events are notifications and may be duplicated or out of order; fetch current resource state before acting when needed.
- OAuth grant, scope, principal, entity, and object policy jointly determine authority.
- Accounting and card workflows have distinct approval, idempotency, sensitive-data, and reconciliation requirements.

## Instructions

1. Define business capabilities, actors, tenants/entities, latency, volumes, data classes, financial effects, recovery objectives, and prohibited actions.

2. Partition separate OAuth applications and services for read-only ingestion, event receipt, accounting writes, and card/spend authority where trust differs.

3. Place an allowlisted typed adapter behind secret-managed token acquisition; centralize rate budget, pagination, error parsing, trace IDs, and schema-version evidence.

4. Ingest verified webhooks into a durable deduplicating queue, process asynchronously, and reconcile current API state plus downstream effects on a schedule.

5. Design deployment, observability, incident isolation, data retention, negative-access tests, and rollback; threat-model every path that can spend, sync, or expose card data.

## Tool Discipline

- Use **Glob** to locate candidate code, manifests, fixtures, and evidence without widening scope.
- Use **Grep** to find relevant endpoints, fields, permissions, identifiers, errors, and stale assumptions.
- Use **Read** to inspect the smallest required local files and authoritative evidence.
- Use **Write** only for a new approved local draft, test, configuration, or evidence artifact.
- Use **Edit** only for a bounded approved change with a known rollback.
- Local file tools do not authorize a Ramp operation or replace owner approval.

## Approval Boundaries

Architecture, security, privacy, business, and finance owners approve boundaries. Ramp approval is required for vendor-restricted surfaces such as production Vault access.

## Output

A context and data-flow design, trust boundaries, application/scope matrix, component ownership, event/reconciliation model, threat analysis, SLOs, rollout, and rollback.

## Error Handling

| Condition | Response |
|---|---|
| One service holds every scope | Split authority by capability and lifecycle, then retest cross-boundary denial. |
| Webhooks directly perform financial writes | Insert a durable deduplicating ledger and reconcile current state before side effects. |
| Observability needs raw payloads | Design field-safe structured signals and restricted evidence access instead. |

## Examples

### Example 1

Design a read-only warehouse exporter with webhook triggers and periodic reconciliation, isolated from accounting-write credentials.

### Example 2

Design an ERP connector whose source ledger, downstream receipt, and Ramp sync receipt form one auditable state machine.

## Validation

- Every component has a single trust purpose, owner, and failure boundary.
- Secrets, card data, identity data, and financial writes follow minimized paths.
- Duplicates, out-of-order events, timeouts, partial failure, and replay are modeled.
- Authority and reconciliation have positive and negative verification.

## Resources

- [Official documentation and contract notes](references/official-docs.md)
- Re-check the dated contract and current OpenAPI schema before any live request.
- Treat unresolved vendor behavior, authority, or financial state as a stop condition.

Files in this skill

  • SKILL.md5 KB
  • references/official-docs.md862 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.