AI agent skill
Verify Frontend
How to verify Massing web/viewer UI changes given that the dev-preview geometry loader stalls. Invoke when you changed apps/web and need to prove it works. Covers typecheck/lint/vitest/build, the tools-panel force-build technique, and honest flagging of flows you couldn't exercise.
·
When to use this skill
Use Verify Frontend when an AI agent needs a reusable SKILL.md workflow for this job: How to verify Massing web/viewer UI changes given that the dev-preview geometry loader stalls. Invoke when you changed apps/web and need to prove it works. Covers typecheck/lint/vitest/build, the tools-panel force-build technique, and honest flagging of flows you couldn't exercise.
When not to use it
Skip Verify Frontend when the task is outside the coding category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.
How to install
- Personal install: create ~/.claude/skills/verify-frontend/SKILL.md (and any bundled scripts) so Claude Code, Claude Desktop, and compatible agents can load it in every project.
- Project install: commit the same folder at .claude/skills/verify-frontend/ so teammates get the skill with the repo.
- Restart the agent session after copying files so it re-scans the skills directory, then ask for the task in words that match the skill description.
What this skill does
# Verify a Massing frontend change
The dev-preview geometry loader **stalls at "preparing geometry"** for all model sizes, which blocks `loadProjectModel` and therefore `buildPanels` (the `panel-tree` model browser). But `buildToolsPanel` (the rail tools) can be forced. Verify what you can; flag what you can't.
## Always ``` cd apps/web && export PATH="/c/Program Files/nodejs:$PATH" npm run typecheck && npm run lint # eslint pinned to 9.39.5; Node 20 (Node 18 breaks build) npx vitest run <path/if/covered> npm run build # strongest compile check; ~1 min ```
## Force the tools rail to build (verifies rail UI live) The preview server is on :5173. In the running preview (via the browser tools), navigate to a project with a source IFC, then: ```js window.dispatchEvent(new CustomEvent('aec:persona', { detail: 'all' })); // forces buildToolsPanel + buildClashPanel // then read #panel-tools — the tool-group sections, buttons, inputs all render even though geometry stalled ``` This verifies: new rail tools/buttons/inputs, the ribbon tabs, the Library palette, the node-canvas launcher, the KEYS shortcut layer (dispatch KeyboardEvents), the Ask/analytical boxes. Exercise handlers by `.click()`ing buttons and reading the result-overlay / DOM.
## What you CAN'T verify live (flag it honestly) - `panel-tree` UI (the Project-Browser spine) — `buildPanels` never runs behind the stall. Verify its DOM-construction logic by reproducing it in the console instead, and say so. - End-to-end calls to NEW API routes against the dev API on **:8093** — that server is often a **stale process** predating your new routes → 404 there even though the client hits the right URL. Confirm via `read_network_requests` that the URL/payload is correct; the backend itself is covered by CI. (Restarting :8093 is the user's call.) - Anything geometry-coupled (placing an element, section cuts, camera).
## Report State exactly what was verified live vs. by typecheck/build vs. not exercised. Never claim an interactive viewer flow "verified" if the stall prevented it.
See memory: `tools-panel-verify-technique`, `web-build-needs-node-20`, `web-eslint-node-pin`, `dev-api-port`.
Intended uses
- Use Verify Frontend when this documented workflow matches the task.
Related skills
Related skills in this directory, for comparison before you install another skill.
coding
Act as a Patient, Non-Technical Android Studio Guide
A reusable prompt for asking an AI assistant to work as Act as a Patient, Non-Technical Android Studio Guide.
coding
Add Ave Record
The main workflow for this repo. Adds one new AVE record end to end.
coding
Add Backend
Guide for adding a backend (Rust or Python) to the agent-sec-core security middleware. Use when creating new backends, integrating Rust or Python code into the security middleware, or extending with new backend actions.
coding
Agent Device
Drive iOS and Android devices for the Expensify App - testing, debugging, performance profiling, bug reproduction, and feature verification. Use when the developer needs to interact with the mobile app on a device.