More screenshots
Works with the AI Agents you already use
Component Stress Lab — See the Edge States
Put a real component's difficult states on a local workbench. Repeat the cases and see which layouts or interactions need attention.
Free
Component Stress Lab — See the Edge States
Example session with this skill installed
Give this Parcel Studio delivery card a local state workbench so I can inspect difficult content and small screens, switch through the states it actually supports, and reset it. Show what you observed; do not fix the product.
- Read your context and instructions
- Compiled the component stress lab
Delivery card workbench — synthetic task result
Built a separate local workbench importing the actual deliveryCard factory and original styles. Product source remains unchanged. Eight fictional cases were rendered at child CSS viewports 768 × 700 and 320 × 700; the parent workbench viewport is recorded separately.
| Case | Actual phase | 768px child | 320px child |
|---|---|---|---|
| baseline | ready | Fits; 768px page | Overflow; 436px page |
| long-copy | ready | Fits; 768px page | Overflow; 436px page |
| long-token | ready | Overflow; 1137px page | Overflow; 1137px page |
| unicode | ready | Fits; 768px page | Overflow; 436px page |
| empty-text | ready | Fits; 768px page | Overflow; 436px page |
| loading | loading | Fits; 768px page | Overflow; 436px page |
| empty | empty | Fits; 768px page | Overflow; 436px page |
| error | error | Fits; 768px page | Overflow; 436px page |
Two layout failures are reproducible: the fixed 420px card extends the 320px page to 436px, and an unbroken title extends the page to 1137px even at the wider viewport. These are observed fixture failures; product code has not been repaired.
Supported loading and empty actions are disabled. The ready action returns the real callback. Error exposes Try again and returns its real callback while remaining in the error phase: the component has no product retry lifecycle, so none is invented. State selection remounts the actual component with supported inputs. Returning to baseline resets exercised local state; the reset audit runs baseline → demanding case → baseline twice.
Isolation is installed before the component import. A deliberately blocked guard probe checks enforcement; inspected resource entries stay on loopback. This proves the observed local run, not all future sessions or unrelated processes.
Deliverables: a local workbench and server, repeatable scenario definitions, source/seam/input/viewport receipts, screenshots, measurements, callback/reset evidence and cleanup instructions. Payment, permissions, cancellation, persistence and a product retry lifecycle are unsupported by this target and remain outside the gallery. RTL product behavior, screen-reader behavior, real asynchronous services and other engines are unverified.
Next action: reproduce the two layout failures through the named scenarios, then authorize a focused product fix if wanted. The package is a workflow for an agent using a project's own tools; this fictional workbench is trial evidence, not a universal gallery runtime.
Connects securely to your tools. The creator never sees your data.
About this skill
A component looks fine with sample text. Then a long title hides its action, an error state never appears in preview, or cached data makes a loading fixture look successful.
Component Stress Lab guides your coding agent to build a local workbench around the actual component. It selects cases from the inputs and branches your code supports, mounts fictional fixtures through the real data seam, and makes each case repeatable.
You get a local preview, a scenario record, and observed failures tied to their inputs, viewports and evidence. The workflow checks that fixture data reached the component and that returning to the normal case resets the state. Unsupported cases and checks that could not run stay visible.
Use it when you want to explore a component's limits or inspect its existing states together. For one known bug, start with a focused reproduction unless you also want broader stress coverage.
What you supply
- An existing component and access to its project source.
- The project's working local runtime and dependencies.
- A coding agent with file-editing and command tools. Rendered verification also needs a browser or the project's browser test tools.
What is included
- The skill instructions.
- Scenario-selection and evidence-record guidance.
- Mounting, local fixture isolation and reset procedures.
- Original-source attribution and MIT license.
This is an agent workflow, not a hosted testing service or a universal gallery generator. It uses your project's tools and existing component API. A finite set of fictional local cases does not establish complete state coverage, accessibility compliance or production reliability. Product fixes and deployment are separate actions.
Example request
“Use Component Stress Lab on our ResultCard. Show its supported loading and error states, long titles at a compact width, and a repeatable return to normal data. Keep all responses local and report what actually ran.”
Worked example
In a retained synthetic Parcel Studio trial, eight cases at two child viewports exposed two reproducible layout failures. Supported states, callbacks and baseline reset were exercised through the real component. Product source stayed unchanged. This is local fictional evidence, not proof of production reliability.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 19 days ago
- Free with an Agensi account
Requires
Permissions
Local project source and fictional fixtures. Browser network access is limited to the local preview; no production service, credentials or external account is required. Creates temporary local workbench files; product fixes need a separate request.