create-verification-skill

Generates a project-local verification skill that drives apps through real user flows.

1|Updated Aug 26, 2026
One-click install
npx skills add https://github.com/edivad1999/stuc-stack --skill create-verification-skill-edivad1999
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: create-verification-skill
Source: https://github.com/edivad1999/stuc-stack/tree/main/skills/create-verification-skill
Command: npx skills add https://github.com/edivad1999/stuc-stack --skill create-verification-skill-edivad1999

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Projects often lack a scripted way to prove that the real app behaves correctly from a user's perspective. This Skill interviews a repository and generates a harness-agnostic verification skill (SKILL.md plus a feature map) that launches the app, drives features programmatically, and captures evidence. ## Core Features & Use Cases - Repo Interview: Detects the app's surface, run command, drive mechanism, observable evidence, and isolation constraints directly from the codebase. - Skill Generation: Writes a canonical docs/verify-<app>/SKILL.md with Launch, Doctor, Drive, Evidence, Cleanup, and Helpers sections, plus thin harness delegates for Cursor, Claude Code, or Codex. - Feature Map Seeding: Creates one file per user-facing feature with a fixed four-heading structure, then proves the generated skill by executing it end to end. - Use Case: An Android project has no scripted UI verification. Run this Skill to generate a verify-<app> skill that assembles via Gradle, drives the app with the android CLI, and captures layout JSON and screen PNGs as proof. ## Quick Start Ask the agent to create a verification skill for this repository so the app can be driven and proven like a real user would.

Frequently Asked Questions about create-verification-skill

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a verification skill for my repository?▼

Invoke /create-verification-skill and the agent interviews the codebase to determine the app's surface, run command, drive mechanism, and evidence options. It then writes docs/verify-<app>/SKILL.md plus a feature map and proves the skill by running it once end to end.

How to verify Android app UI behavior with the android CLI?▼

The generated skill assembles the APK with Gradle, deploys it with android run, then drives it using android layout, screen resolve, and screen capture. Evidence includes layout JSON, screen PNGs, and journey JSON stored in a named artifacts directory.

Does this work for non-Android apps like web or CLI projects?▼

Yes, the generator is harness-agnostic and supports web UIs, CLIs, TUIs, desktop apps, APIs, and libraries. It reuses existing Playwright or Cypress specs, expect scripts, PTY helpers, or curl-able endpoints as the drive mechanism.

What is the difference between the canonical skill and harness delegates?▼

The canonical skill and feature map live under docs/verify-<app>/ and hold all Launch, Doctor, and Drive instructions. Harness folders like .cursor/skills or .claude/skills only contain a thin delegate that registers the slash command and points back to the docs skill.

Why does the generated skill need a proof run before handover?▼

A generated skill that was never executed is treated as a draft, not a deliverable. The proof run launches the app, runs doctor, drives one mapped feature, captures evidence, and confirms cleanup does not delete the proof artifacts.

What are the limitations of host unit tests as verification?▼

Host unit tests run through Gradle do not count as device proof of UI behavior. The skill requires driving the real user path on a device or emulator and capturing observable state such as layout JSON or screenshots.