- Home
- Skills
- Testing & Debugging
- Web Form Failure & Recovery Test Kit
Web Form Failure & Recovery Test Kit
Check one local JSON form for false success, lost input, wrong payloads, excessive in-flight requests and failed manual retry using a supplied mocked-browser checker.
Free
Works with the AI Agents you already use
Web Form Failure & Recovery Test Kit
Example session with this skill installed
Synthetic example: test my approved localhost enquiry form with fake values. Declare the exact JSON POST, success/error feedback, preserved fields, maximum one in-flight request and permitted manual retry. Use the supplied checker without editing the app or sending a real enquiry; report blocked setup separately from observed violations.
- Read your context and instructions
- Compiled the web form failure
Illustrative synthetic outcome: report.json and report.md identify server_rejection as failed because success feedback appeared after a mocked 503. Include reproduction, contract/runner hashes, actual runtime and omitted behavior. persistenceVerified: false; duplicateBusinessEffects: null. This is not evidence of backend persistence or a real customer result.
Connects securely to your tools. The creator never sees your data.
About this skill
A green success message is not proof that your form sent the right data.
For developers and agencies reviewing a local enquiry, profile or settings form. Declare its expected behavior and run five focused browser checks: accepted response, rejected response, transport abort, repeat click while pending, and retry after rejection. Version-2 contracts can add a native-invalid-input check without treating an expected zero-request validation result as a missing-capture defect.
Includes an agent workflow, a standalone Node.js checker, a contract reference and clearly synthetic v1/v2 examples. Version 2 supports native fill/select/check controls, exact typed nested JSON expectations and a shared initially empty status container. Reports keep input/runner hashes, actual runtime and per-case pass/failure/blocked observations. The checker does not edit your app or silently fix its tests.
Requires Node.js 22+, an existing approved playwright-core runtime and installed Chrome. This prototype supports one loopback JSON POST form with stable pre-existing feedback containers; mapping and separately approved browser execution are necessary.
This is a free early release, version 0.2.1. Pinned component trials and source/ZIP checks do not establish general real-project or marketplace delivery compatibility. File uploads, custom widgets, conditional feedback mounting, multiselect/radio groups and login setup remain outside this runner.
Responses are mocked in the browser: no database persistence, email delivery or duplicate business effects are verified. Multiple requests may be legitimate under your contract. Missing setup/expectations remain blocked, not a clean bill of health. This is not a full web/security audit, production scanner, project-specific test generator or marketplace compatibility guarantee.
Independent buyer usefulness, comparative benefit, demand and Agensi script delivery are not established. No saved-time or sales promise is made. Public test tooling such as Playwright already offers free general-purpose automation; this package's proposed value is a bounded failure/recovery contract and repeatable checks.
Early 0.2.1 release. Loopback JSON POST only; five cases and optional native-validation case. No file uploads, custom widgets, multiselect/radio groups, iframes, login setup, conditional feedback mounting, hosted targets or arbitrary JavaScript hooks. No app changes, real-backend/persistence/business-effect verification, complete web/security audit or guaranteed absence of later actions. Independent usefulness and platform delivery remain unverified.
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 20 days ago
- Free with an Agensi account
Requires
Permissions
Allowed Hosts
File Scopes
Browser execution, clicks and traffic interception require explicit permission for the exact loopback origin. Declared POST responses are mocked; unexpected writes, external HTTP and WebSockets are blocked. Browser routing is not an OS/network sandbox. No app edits, dependency installation, real enquiries, credentials, input upload or production testing are authorized by default. Model-provider data handling applies separately.