---
title: "Building secure mobile apps in the agentic era (2026) - Shopify"
description: "Shopify’s mobile app team now ships 11x more code per week than a year ago, averaging 434,000 lines changed per week during a 12-week rewrite. To maintain security at this pace, they built a system that maintains a map of each app’s attack surface, organized into surfaces such as deep links, WebViews, and stored data. Each surface has a skill derived from OWASP MASVS/MASTG and lessons from past incidents, which is used to inventory and review one item at a time. A daily pipeline compares the latest inventory with the last accepted map, reviews each changed item with read-only code access, and publishes a new map and changes page. Agents also monitor security news for industry-wide vulnerabilities and check them against the map and code, posting briefings only when the app is affected."
---

# Building secure mobile apps in the agentic era (2026) - Shopify

[Shopify](https://yomu.fyi/company/shopify) · Michael Hughes · Nov 18, 2026

**Type:** Explainer

## Summary

Shopify’s mobile app team now ships 11x more code per week than a year ago, averaging 434,000 lines changed per week during a 12-week rewrite. To maintain security at this pace, they built a system that maintains a map of each app’s attack surface, organized into surfaces such as deep links, WebViews, and stored data. Each surface has a skill derived from OWASP MASVS/MASTG and lessons from past incidents, which is used to inventory and review one item at a time. A daily pipeline compares the latest inventory with the last accepted map, reviews each changed item with read-only code access, and publishes a new map and changes page. Agents also monitor security news for industry-wide vulnerabilities and check them against the map and code, posting briefings only when the app is affected.

## Context

Shopify’s mobile app team ships 11x more code per week than a year ago (averaging ~434,000 lines changed per week during a 12-week rewrite). The industry is seeing record numbers of vulnerabilities even in human-reviewed codebases. The team needed a way to keep shipping quickly while minimizing attack surface and enforcing security practices proactively.

## Approach / What changed

A scheduled daily pipeline that inventories the app’s attack surface (deep links, WebViews, stored data, permissions, dependencies, etc.), compares it with the last accepted map, reviews each changed or new item one at a time with read-only code access, and publishes a new map and a changes page. Skills written for each surface encode best practices from OWASP MASVS/MASTG and lessons from past incidents. A separate agent monitors security news and checks each item against the map and code.

## Takeaways

- Maintain a map of each app’s attack surface grouped into surfaces (deep links, WebViews, stored data, permissions, dependencies) with a verdict and code link for each item.
- Write security knowledge as skills derived from OWASP MASVS/MASTG and past incidents; skills are not tied to one tool and can be loaded by any agent working in the code.
- Review one item at a time with only the code that item touches, and use frontier cybersecurity models for questions about exploitability.

## FAQ

### How does the review process avoid vague results?

Each skill first takes a complete inventory of its surface (e.g., every deep link route and parameter, every WebView and host, every Keychain entry). Then the model reviews one item at a time, with only the code that item touches. For a deep link, that means looking at one route, every parameter it reads, and the code that reads each one, asking what happens when a parameter is hostile.

### How are lessons from past incidents incorporated?

When an incident or HackerOne report is handled, the team fixes the bug and then adds that class of problem to the skills. Every subsequent run checks the current code against all of those classes, and each surface has a short checklist of the best practices expected.

### How does the system handle vulnerabilities outside the repo?

Agents regularly read security news and check each item against the map and the code. The agent posts a short briefing and only pings the team when something affects the app. The same skills are also used on third-party libraries before adoption.

**Tags:** [Agentic AI](https://yomu.fyi/topic/agentic-ai), [Attack Surface](https://yomu.fyi/topic/attack-surface), [Shopify](https://yomu.fyi/topic/shopify)

- Source: [Shopify](https://shopify.engineering/secure-agentic-apps)
- Source URL: https://shopify.engineering/secure-agentic-apps
- Ingested by Yomu: 2026-11-19T03:00:15.935Z

[Read original post](https://shopify.engineering/secure-agentic-apps)
