Ecommerce Process Mining — Development skill for Claude Code

ecommerce-process-mining

[](https://agentskills.io) [](https://www.skills.sh) [](https://docs.claude.com/en/docs/claude-code/skills) [](https://developers.openai.com/codex/skills) [](https://github.com/google-gemini/gemini-cli/blob/main/docs/cli/skills.md)

An [Agent Skill](https://agentskills.io) that teaches an agent to build employee process mining for an e-commerce back office in a **Next.js App Router** app: a consented Manifest V3 browser extension that reads DOM events rather than pixels from the admin tools staff already use, a gap-capture form for the work a browser cannot see, an ingest route that scrubs before it stores, and a batch AI pipeline that turns the result into per-role SOPs and a ranked automation shortlist.

**In e-commerce the screens are full of other people's data.** The employee consents; the customer whose order, return or ticket is on screen never did, and never can be asked. That single fact decides the design: consent is re-read on every batch rather than cached, exclusions run in the extension and again on the server, values that would have to be scrubbed are never read at all, screenshots default to off because the scrubber reads JSON and not pixels, and managers get suppressed aggregates instead of rows. Capture is a privacy system that happens to produce data.

This skill was written by the engineer who has shipped this module. The earlier implementation it was audited against was the employee process-mining module of a back-office platform, of which the consent record, the gap-capture form and a manager view ran while the capture path existed as design notes. The temp