Choose a small defect with an observable result, then document it so another person can repeat the same actions and compare the outcome with a stated requirement. A fictional form that rejects an allowed input is enough to demonstrate clear technical communication, provided you label the scenario as fictional and distinguish an illustrative report from a test you actually performed. The goal is to help a remote teammate work without needing an immediate call to fill in gaps. A strong sample states the starting conditions, exact input, steps, expected and actual behavior, supporting evidence, and limits. It does not need a dramatic failure or a claim that you solved a client’s problem. What should you put in the demo? Use a harmless sign-up form with one display-name field and a submit button. Give the form an explicit fictional requirement: after leading and trailing spaces are removed, a display name containing two or more characters should be accepted. Invent a defect in which the form incorrectly rejects a two-character name and displays “Display name must contain at least 3 characters.” This narrow scenario gives the reviewer a clear comparison: the stated minimum is two characters, but the displayed validation message requires three. It avoids real accounts, customer information, and security-sensitive behavior. If you build the demo, keep its files and any test data separate from an employer’s systems. If you do not build it, present the report as an illustrative exercise rather than evidence of a completed test. MDN’s guidance on reporting browser bugs recommends a minimal, standalone test case and details such as the browser version, expected and actual results, and screenshots. Those principles also make an application-level work sample easier to review. The example below describes a fictional application defect; it does not establish a browser defect. What does a reproducible report look like? The following is a hypothetical report. Its environment and result are invented to show the level of detail to capture; replace them with your own observed details if you build and test the demo. Title: Display-name field rejects a two-character name despite the stated minimum. Environment: Fictional sign-up demo, version 1.0; desktop browser and operating system versions recorded by the tester. Preconditions: Open a fresh copy of the demo with an empty display-name field. The demo requirement says a trimmed name of at least two characters is valid. Steps: 1. Enter “Li” in the display-name field. 2. Select Submit. 3. Observe the message beside the field and whether the demo displays its success state. Expected: The form accepts “Li” and displays the demo success state because the name has two characters. Actual, in this fictional scenario: The form displays “Display name must contain at least 3 characters” and does not show the success state. Impact: A person using a valid two-character display name cannot complete this demo form. Evidence to include if you run the test: a screenshot showing the entered value and the validation message, plus the demo files or a self-contained test case a reviewer can run. Limits: This example makes no claim about other name lengths, browsers, devices, or the underlying cause. No real user or customer data is involved. A reviewer should be able to identify the requirement, input, action, and discrepancy without interpreting a long narrative. If your real test differs from the hypothetical result, report what happened rather than preserving the example’s expected defect. Report element | Reproducible version | Vague version Title | Two-character display name is rejected | Sign-up is completely broken Input and action | Enter “Li” and select Submit | Try signing up Result | A three-character minimum message appears | The form fails Scope | One demo input and stated environment | Claims every user is affected How do you make the steps repeatable? Start with a known state. Say whether the form is empty, whether a previous attempt needs to be cleared, and which version of your demo the reader should open. Give actions in order and use the labels visible on the form. If a specific value matters, include it exactly; “enter a short name” leaves too much room for interpretation. Record the browser and operating system versions you actually used, along with any demo version or revision that identifies the tested files. Do not fill in plausible versions from memory. If you cannot establish a detail, say it is unknown. The purpose is to help someone recreate the conditions, not to make the report look more technical. Run your own instructions from the stated starting point before sharing the sample. If the outcome changes between attempts, describe that variation and how many attempts you made rather than calling the defect consistent. A second person following the steps can reveal missing instructions, but do not claim independent verification unless it happened. Keep the example small. A separate navigation journey, decorative fields, and unrelated error messages make it harder to identify which action matters. MDN recommends reducing a test case to the relevant behavior; for this exercise, that means keeping the display-name rule and its visible result. What counts as evidence, and what does it prove? A screenshot can show the value and message together, while a short recording can show the sequence leading to the result. Either is supporting evidence for the report; neither replaces written steps. Make the important field and message readable, and check the whole image for private information before sharing it. A runnable demo gives a reviewer a way to attempt reproduction. Include brief instructions for opening it and identify the specific version the report describes. If the sample includes only a written scenario, say so: the reader can assess your reporting structure but cannot independently reproduce the behavior from the report alone. Separate observation from explanation. “The form displays a three-character minimum message” describes an observable result. “The validation function uses the wrong comparison operator” is a hypothesis unless you inspected the code and have evidence for it. You may suggest a possible investigation, but label it as a hypothesis and avoid presenting it as the confirmed cause. A single successful reproduction also has a limited scope. It does not establish how many people would be affected, whether the issue occurs in other browsers, or whether a production application has the same problem. List additional checks as future work unless you actually perform them. Why is a calm report stronger than a dramatic one? Compare “Sign-up is completely broken and nobody can register” with the sample’s narrower title and result. The dramatic version suggests a much wider failure but provides no input, requirement, or visible outcome. A teammate must ask basic questions before deciding what to investigate. The concise version gives the next person a starting point and a reason to care: the fictional requirement permits a name that the form rejects. Its impact statement follows from that scenario without assigning a business loss or severity rating. If a team uses its own severity scale, supply the facts needed for that team to assess priority rather than inventing a classification. There is a trade-off. A very short report can omit essential conditions, while a long report can bury the failure. Keep the main reproduction path near the top; put additional attempts, context, and open questions after it. Include a failed or inconclusive check only when it helps define the boundary of what you observed. How should you present the sample in an application? Introduce it in one sentence as a fictional exercise, then say what the reader will find: a requirement, a reproduction path, the expected and actual results, evidence if available, and stated limits. If you built the form yourself, explain that briefly and accurately. If you only wrote the hypothetical report, make that clear too. Neither should be described as client work. Match the presentation to the role. For a quality assurance application, emphasize the test conditions and the checks you would run next. For a developer role, you might add a clearly labeled hypothesis or a proposed regression test: under the fictional requirement, submitting “Li” should reach the success state. Keep proposed work distinct from tests you actually ran.
Talyca also has a separate guide on entry-level remote jobs in Canada, including roles that suit applicants with limited experience.
Use the role description to decide which parts of the sample to emphasize; the report itself should remain understandable to someone who has never seen the demo. Before sharing, make one final pass: can the reader find the exact input and starting state, tell the requirement from the observed result, open any supplied evidence, and see what was not tested? If any answer is no, revise that part before adding more detail. Sources • When and how to file bugs with browsers • Creating a support ticket • Source for: Reproduction steps, expected results, exact error messages, and screenshots provide useful context i