create-verification-skill

Generates a project-local verification skill that drives an app like a user.

Updated Sep 23, 2026
One-click install
npx skills add https://github.com/lucasheriques/pstack --skill create-verification-skill-lucasheriques
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: create-verification-skill
Source: https://github.com/lucasheriques/pstack/tree/main/skills/create-verification-skill
Command: npx skills add https://github.com/lucasheriques/pstack --skill create-verification-skill-lucasheriques

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Projects often lack a scripted way to prove that real user-facing behavior works. This Skill interviews a repository and generates a tailored verification skill (.claude/skills/verify-<app>/) that launches the app, drives it the way a user would, and captures evidence. ## Core Features & Use Cases - Repo Interview: Identifies the app's surface (web UI, CLI/TUI, API, desktop), run command, driving harness (Playwright, CDP, tmux/PTY, HTTP), observable evidence, and isolation constraints directly from the codebase. - Skill Generation: Writes a complete SKILL.md with Launch, Doctor, Drive, Evidence, Cleanup, and Helpers sections grounded in real selectors and commands from the repo. - Feature Map Seeding: Creates a features directory with one file per user-facing feature, following a strict four-section contract (Sub-features, user POV entry points, harness driving steps, Gotchas). - Self-Proof: Runs the generated skill end to end once — launch, doctor, drive one feature, capture evidence, clean up — before handing it over. - Use Case: Point it at a repo with no scripted UI testing, and it produces a working verification skill plus a feature map that later agents can execute cold. ## Quick Start Ask the AI to create a verification skill for this repository so its user-facing behavior can be driven and proven automatically.

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 the skill on your repo and it interviews the codebase to find the app's surface, dev command, and driving harness. It then writes `.claude/skills/verify-<app>/` with launch, doctor, drive, evidence, and cleanup instructions tailored to what it found.

What kinds of applications can this verification approach drive?▼

It supports web UIs, CLIs and TUIs, desktop apps, APIs, mobile apps, and libraries. It prefers existing harnesses like Playwright or Cypress specs first, then falls back to generic recipes such as browser/CDP for web, tmux/PTY for terminals, or plain HTTP for services.

Does the generated verification skill get tested before delivery?▼

Yes. The skill runs its own instructions end to end once: launch, doctor, drive one mapped feature, capture evidence, and clean up. A generated skill that was never executed is treated as a draft, not a deliverable.

What is the feature map in a generated verification skill?▼

The feature map is a `features/` directory with a README index and one file per user-facing feature. Each file lists sub-features, user entry points, exact driving commands with observable results, and gotchas, serving as the repo's maintained verification source.

Can two verification instances run side by side?▼

Only if the app supports isolation through separate ports, data directories, or profiles. If not, the generated skill says so explicitly, because double-driving a shared instance risks corrupting the user's session.

What evidence does a verification run capture?▼

It captures the user action and resulting state, such as screenshots, ARIA snapshots, terminal transcripts, response bodies, logs, and exit codes. Proof artifacts survive cleanup in a named location, and side effects like files written or rows inserted are verified too.