- Home
- Skills
- Testing & Debugging
- StarIA Form Flow QA
More screenshots
Works with the AI Agents you already use
StarIA Form Flow QA
Run configured web form checks with Playwright. Review local HTML, JSON and Markdown reports, masked screenshots and explicit PASS, FAIL and NOT TESTED results. Verified Linux reference environment.
$19
StarIA Form Flow QA
Example session with this skill installed
Test only the supplied synthetic local form. Use the deliberate-defect demo with Chromium, keep traces disabled, write to a new report directory and return every scenario result, including failures and NOT TESTED checks. Do not install dependencies without approval or visit external targets.
- Read your context and instructions
- Compiled the staria form flow
- Generated the UI component
Actual local acceptance result — native defective fixture
StarIA Form Flow QA 0.26.0rc4, Chromium, recorded on 7 November 2026. Synthetic local target; address withheld. Runtime error: none. Duration: 9.1 seconds. Screenshots redacted, traces disabled.
7 PASS · 2 FAIL · 5 NOT TESTED
| Scenario | Actual result | Observed POSTs | HTTP statuses |
|---|---|---|---|
| required_fields | PASS | 0 | none |
| invalid_email | PASS | 0 | none |
| happy_path | PASS | 1 | 201 |
| duplicate_submit | FAIL | 2 | 201, 201 |
| server_error_recovery | FAIL | 1 | 503 |
| network_error_recovery | PASS | 2 | 201 |
| persistence | PASS | 1 | 201 |
| keyboard_flow | PASS | 1 | 201 |
| mobile_layout | PASS | 1 | 201 |
| controls_roundtrip | NOT TESTED | — | — |
| conditional_reveal | NOT TESTED | — | — |
| wizard_navigation | NOT TESTED | — | — |
| dependent_choices | NOT TESTED | — | — |
| accessibility_names | NOT TESTED | — | — |
Each PASS verified its configured assertions against observed browser behavior. Both FAIL results were classified ASSERTION: a configured form assertion failed and requires review of the scenario and locally masked evidence. The five NOT TESTED scenarios were not selected or reached. Console errors were observed during server and network recovery; details were withheld in the report.
Local artifacts: report.html, report.json, report.md and before/after masked screenshots for each executed scenario. No trace was captured. The full scenario matrix above is a readable transcription of the stored acceptance report, not a new execution or an Agensi MCP session.
Next steps: inspect the two deliberate fixture defects, retain FAIL in the report, review masks before sharing and compare with the supplied healthy demo. PASS applies only to executed assertions; unexecuted checks are not approved. No real customer data, external site, login or payment was used.
report.html
TSX · React component
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Browser evidence for web forms
StarIA Form Flow QA runs configured signup, contact and dynamic form checks with a proprietary Playwright runner. Give your agent reviewed selectors, synthetic inputs and expected outcomes, then inspect local HTML, JSON and Markdown reports with masked screenshots and explicit PASS, FAIL and NOT TESTED results.
Who it is for
Developers, QA professionals, freelancers and individual professionals in product teams and small agencies who need repeatable form checks before delivery. Agensi purchases are personal and nontransferable; this listing does not grant a shared team licence.
What you can check
- Required fields and invalid email, including whether a POST was prevented.
- Successful submission and duplicate-submit prevention.
- Controlled network/server error recovery and configured persistence.
- Keyboard order, focus traps and a mobile viewport.
- Native values, conditional fields, steps and dependent choices.
- Accessible names for configured controls.
There are 14 configurable scenarios. Run applicable scenarios; the rest remain NOT TESTED. Local acceptance fixtures used native HTML, React Select, Material UI, React Hook Form, Formik and Zod. Framework names do not guarantee compatibility with your component or version; selectors and expectations require review.
Delivery and first use
Delivery is a textual skill with its unchanged Python entry executed temporarily by an Agensi-connected agent. Python, Playwright and browser binaries are separate prerequisites, installed only with your permission. No StarIA account or LLM API key is required by the runner. Your agent provider has its own requirements and privacy policies.
Ask your agent to run the supplied doctor for Chromium, then the healthy local demo to a new report directory and the deliberate-defect demo to another new directory. Healthy native demo: 9 PASS, 0 FAIL, 5 NOT TESTED. Defective demo: 7 PASS, 2 expected FAIL, 5 NOT TESTED. Keep the failures visible. Reusing a report directory is refused to preserve earlier evidence.
Verified environment
Version 0.26.0rc4. Ubuntu 26.04.1 LTS x86_64, Python 3.14.4, Playwright 1.64.1 and its browser binaries. Reference acceptance runs used Chromium, Firefox and WebKit on Linux. Windows/macOS are not homologated. The temporary launcher uses POSIX process-group supervision. Mobile viewport testing is not physical-device or native Safari certification.
Limits, privacy and operational safety
Configured assertions only. No automatic bug fixing, universal compatibility, full accessibility audit, security certification or LGPD/WCAG compliance claim. A FAIL can indicate an application issue, selector drift or an execution problem. PASS approves only the executed assertion; NOT TESTED is not approval.
Reports are generated locally and are not automatically uploaded to StarIA. Screenshot masks are best effort; review images before sharing. Traces are off by default and may contain data. HTTP/WebSocket guards are mitigations, not universal network isolation. Forced executor loss or SIGKILL may prevent temporary-source cleanup.
Use only explicitly authorized targets and synthetic data. Do not execute real payments, customer logins, destructive production submissions or sensitive-data workflows through the examples. External staging requires explicit authorization.
Licence, support and updates
StarIA code and instructions remain proprietary. Access follows Agensi's personal, nontransferable licence and current purchasing rules. Published approved updates are supplied through Agensi without a separate update charge. No guaranteed release schedule or lifetime maintenance is promised. Agensi's official guarantee applies. Setup support: 30 calendar days of asynchronous assistance through Agensi Messages with the documented Linux setup, supplied configurations and report interpretation. No SLA or custom integration work is included.
Installation and first run
On the verified Linux reference, create a Python virtual environment, install playwright==1.64.1 and the selected Playwright browser binaries with your permission. Ask the agent to run the supplied temporary entry with --doctor --browser chromium (or the selected supported browser). Continue only when the doctor is ready. The agent runs the supplied scripts unchanged and removes temporary skill sources after use; keep your reviewed configuration and generated reports.
Reference acceptance evidence was recorded on 7 November 2026. The exact standalone buyer entry was also rechecked on 8 November with Firefox in an isolated environment: the healthy and deliberate-defect local demonstrations produced their expected outcomes. Restricted executors may block Chromium or lack WebKit OS libraries; unexecuted checks remain NOT TESTED. Live Agensi MCP execution is not part of that local evidence.
Linux reference environment only; Windows/macOS and physical mobile devices are NOT TESTED. Requires reviewed selectors, explicit expectations and Python/Playwright/browser prerequisites. Configured form assertions are not a full accessibility/security audit or compliance certification. FAIL requires triage; NOT TESTED is no verdict. Masking is best effort; traces can expose data. Network guards are mitigations, not universal isolation. SIGKILL/executor loss can prevent cleanup. Restricted executors can block Chromium or lack WebKit OS libraries. Local acceptance evidence does not certify live Agensi MCP execution.
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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 8 days ago
Requires
Permissions
Reads reviewed local configuration and synthetic upload fixtures. Writes local reports and masked screenshots to a new output directory. Runs temporary own Python sources and Playwright browser processes. Only explicitly authorized URLs; bundled pilots use loopback HTTP. No StarIA service, account, API key or telemetry upload. Browser networking is not a universal isolation boundary. Dependency installation requires separate consent.