skills.rest
Neko-Catpital-Labs
Neko-Catpital-LabsOFFICIAL
Explore all skills by this author →
GitHub

verify

Verify Invoker desktop UI changes by running Playwright specs, visual proofs, and owner queries.

18|5|Updated Apr 17, 2026
Search Tags:#verification#playwright#ui testing#electron#e2e testing#feature map#visual proof
One-click install
npx skills add https://github.com/Neko-Catpital-Labs/Invoker --skill verify-neko-catpital-labs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: verify
Source: https://github.com/Neko-Catpital-Labs/Invoker/tree/main/skills/verify
Command: npx skills add https://github.com/Neko-Catpital-Labs/Invoker --skill verify-neko-catpital-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @playwright/test, and includes scripts (resource) and references (resource) components.

What problem does it solve? Agents working on the Invoker desktop app often claim UI changes are done without evidence. This Skill closes that loop by routing every UI or live-path claim through deterministic verification commands instead of asking a human to check. ## Core Features & Use Cases - Feature-mapped proving: A catalog under references/features/ maps each UI surface (command palette, workflows, merge gate, terminal drawer, and more) to a concrete Playwright spec or repro command, run via prove <feature-id>. - Environment doctor and drift gate: doctor fails closed on stale UI/app builds and missing Playwright, while catalog --check enforces that required sidebar testids and prove paths still exist. - Isolated Electron driving and visual proof: Drive commands (screenshot, click, press, send) launch an isolated Electron build — never the user's open window — and visual-proof wraps capture-before/after/compare for PR pixel evidence. - Use Case: After changing the command palette behavior, run doctor, then prove command-palette, capture a screenshot, and only then claim the change is verified. ## Quick Start Ask the agent to verify the Invoker command palette change by running the verify skill's doctor and prove command-palette commands and showing the screenshot evidence.

Frequently Asked Questions about verify

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

FAQPage Schema
How do I verify an Invoker UI change before claiming it is done?▼

Run control-invoker doctor to confirm builds are fresh, then run prove with the matching feature id from references/features, such as prove command-palette. For pixel changes, capture visual proof and open the screenshot yourself before claiming done.

How do I run a Playwright spec for a specific Invoker surface?▼

Use node skills/verify/control-invoker.mjs prove <feature-id>, which reads the prove command from that feature's frontmatter and executes it from the repo root. Add --dry-run to see the resolved command without running it.

Does the verify skill attach to my open Invoker desktop window?▼

No. Drive commands launch an isolated Electron build with a temporary user-data directory, test database, and its own IPC socket. Attaching to the user's already-open window is explicitly forbidden.

Why does control-invoker doctor fail before running a prove command?▼

Doctor fails closed when packages/ui or packages/app dist builds are missing or older than their src trees, when Playwright is not installed, or when the feature map directory is absent. Rebuild the stale package or run pnpm install first.

What is the catalog --check command used for?▼

catalog --check is a drift gate that verifies every feature file has a prove command pointing at an existing path and that all required sidebar testids are covered by at least one feature. It fails if the feature map has drifted from the codebase.

When should I not use the verify skill?▼

Do not use it for non-UI changes that have no Invoker surface impact; the stays-silent fixture defines that case. It also does not rewrite Playwright specs, replace merge-gate visual-proof capture, or attach to live sessions.

Related Skills

Explore More Software Engineering Skills→
podman-desktoppodman-desktop
8.0k

storybook-hmr

Applies a local git patch enabling Storybook hot reload for UI component source changes.

Official
Intermediate
vTRKAvTRKA

pre-pr-check

Verify pull request diffs against intent and produce evidence bundles.

Community
Advanced
yongwoonyongwoon
8

ywc-toolkit-eval

Evaluate Claude Code skills and agents with scoring and prioritized backlogs.

Community
Advanced
Kiln-AIKiln-AI
5.0k

playwright

Drive Kiln's web UI in a real browser and run its Playwright end-to-end test suite.

Official
Advanced