More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    The Inspector: QA validation page with evidence

    1

    Your agent reads the tickets and the code, and gives your tester one offline page with the exact expected text, pass/fail buttons, pasted screenshots and ready tracker comments.

    $9

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    The Inspector: QA validation page with evidence

    The Inspector: QA validation page with evidence

    Example session with this skill installed

    The two tickets in tickets.md (TS-41 password reset, TS-57 cart discount codes) are on staging. Build me a validation page with the exact texts I should see.

    • Read your context and instructions
    • Compiled the the inspector

    Wrote qa-ts41-ts57.html: 2 ticket(s), 15 scenario(s), 1 need confirmation.

    QA validation for TS-57: FAIL
    1 passed, 1 failed, 0 skipped, 5 not run (7 scenarios).
    Tested by Ana on 2026-11-17 12:47.

    FAIL: SAVE10 takes 10% off a $50.00 subtotal
    Expected: Code SAVE10 applied.
    Expected: Discount (SAVE10): -$5.00
    Note: Notice shows, but there is no discount line and the total stays $55.00.
    Screenshots: 1

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Trusts nothing without a screenshot.

    Manual QA loses time in small ways. The tester checks a label against memory. The ticket says one thing and the code says another. Screenshots end up on the desktop, and the ticket comment says "tested, works".

    The Inspector splits the job. The agent reads, the human tests.

    1. The agent collects the ticket text from what you give it: pasted text, a file, gh issue view, or a tracker tool you already have connected (read only).
    2. It reads the code or the PR diff and copies every expected string and value character for character, with the file:line it came from. When it has to guess, it marks the scenario "needs confirmation" instead of inventing a value.
    3. A small Python script checks the scenarios file and renders one self-contained HTML page.

    The tester opens the page next to the app. Each scenario has steps, the exact expected text with a copy button, PASS / FAIL / SKIP, a notes box and screenshot evidence pasted straight from the clipboard. Progress shows at the top. Marks survive a reload.

    What you get at the end:

    • A plain-text validation comment per ticket, ready to paste into Jira, Linear or GitHub.
    • A bug report draft for every FAIL: title, environment, steps, expected, actual, evidence.
    • Export to JSON (with screenshots) or Markdown, and a summary command that turns the export into a Markdown report and image files.

    The page works offline: no CDN, no server, no network calls, and all ticket text is escaped. The agent never posts to your tracker or changes a ticket without your explicit yes.

    For developers who test their own tickets, QA engineers, and small teams without a test management tool.


    Part of The Guild, agent skills by Sesisnando Neto: The Oathkeeper (sealed acceptance checks), The Quartermaster (quests and decisions), The Tracker (root-cause proof), The Archivist (memory hygiene), The Lorekeeper (repo principles from history).

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.

    Results live in one browser. Another browser or a cleared profile starts empty, and the page cannot import an export. Clipboard image paste depends on the browser; use Choose file or drag and drop if paste is blocked. Expected values are only as good as the code the agent can read; text from a CMS, a database or a third party needs confirmation. The comment is plain text and screenshots are attached to the ticket by hand. The page does not record video or capture network traffic.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    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 today

    Software
    Python 3.8+A modern browser
    Optional
    Git
    orthe gh CLI (to read PRs and issues)

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    File Scopes

    scenarios.json
    and
    qa-*.html
    in
    the
    project
    reads
    code
    ticket
    files

    No network access. The script uses the Python standard library only. Tracker tools, if present, are used read-only unless you approve a specific write.

    Listedtoday

    What's inside

    Frequently Asked Questions

    No. The script only reads the scenarios file and writes an HTML file. The page has no network calls and blocks them with a Content Security Policy. Results stay in your browser until you export them.

    No. The agent prepares the page. A person runs the tests and decides PASS or FAIL. The agent drives a browser only if you ask it to.

    Not on its own. The page builds the comment text for you to paste. The skill tells the agent never to post, attach or change a status without your explicit yes for that action.

    Any. The agent reads tickets from pasted text, files, the gh CLI, or a tracker tool you already have connected. The comment is plain text, so it pastes anywhere.

    In your browser's IndexedDB for that page. Export JSON keeps them, and the summary command can write them out as image files.

    Every expected value carries its source (file:line or ticket). Values the agent could not confirm show a yellow warning, so the tester checks with the owner before marking FAIL.

    New in Testing & Debugging