CatalogueDelivery Documentation & QA
End User Manual
Walks the running application and writes the manual the people who use it actually need.
What it examines
- The screens staff use: lists, a record, and the form that creates one
- Branching cases, such as a person who already exists
- The words on the screen, in plain language
- Screenshots from the running app, retaken when the app changes
What you get back
An end-user manual with screenshots, written for staff rather than for engineers.
Dependencies
- An AI coding agent that reads SKILL.md files, such as Claude Code, Codex or Gemini CLI
- Browser control for your agent (these skills are written for the claude-in-chrome tools), so the agent can drive and screenshot a real browser
- Google Chrome, signed in and able to reach the app
- A genuinely running copy of the app, with a test login. The skill documents the real screens, so a mock-up will not do.
- The backend repository, if a business rule needs checking against its spec
- Optional: pdftoppm, to check the finished PDF
How it works
It walks through the running app in Chrome as a staff member would: dashboard, lists, a record’s detail view and the create form. It takes real screenshots as it goes.
It also documents the key scenarios you name, such as a record that already exists versus a new one. It does not invent scenarios.
It writes plain-language steps for the client’s own staff, with screenshots embedded in an HTML template, then renders a PDF with headless Chrome.
Setup
- Unzip the download. Copy the skill folder into your agent's skills folder (for Claude Code, .claude/skills/ or ~/.claude/skills/).
- Start the app and its backend, and have a test login ready. Set up browser control for your agent (the claude-in-chrome extension, if you use Claude Code) and sign in to Chrome.
- Restart your agent, then ask it to run the generate-user-manual skill.
- Tell it which client or workflow to cover and which scenarios need their own walkthrough.
Run it with your agent
Copy the prompt, fill in the parts in square brackets, and paste it into your agent.
Run the generate-user-manual skill. The app is running at [URL], and I will sign in with [test login]. Client or workflow to cover: [name]. Scenarios that need their own walkthrough: [list].