Practice tasks

Practice on realistic client briefs

Practice tasks are realistic client scenarios with source material, deliverables, and review criteria. They are not paid work.

Task filters

Search by keyword, status, or list

Clear

Showing: 10 open practice tasks. Practice tasks stay reusable for multiple contributors; closed practice records are in History.

Practice TaskResearchOpen

Practice: Complete an AI customer-support tool research report

Source material: A cross-border ecommerce team has a rough list of AI customer-support tools and needs a client-readable comparison report before choosing what to test.

Submit a comparison table, source fields, use-case fit, risks, recommendation order, and open questions for the client.Must cover the supplied tools; must separate verified facts from items to verify; must not invent pricing, features, or customer cases.
Practice TaskAutomationOpen

Practice: Compare three given automation options

Source material: A small business wants form submissions written to Google Sheets, a Feishu group notified, and a weekly summary generated. Compare Zapier, Make, and a custom Node.js script.

Submit an option comparison, recommendation order, fit conditions, failure risks, monitoring suggestions, and client questions.Must compare all three supplied options; must explain the recommendation in terms a small business can understand; must not answer only 'it depends'.
Practice TaskCodingOpen

Practice: Explain and triage a provided code snippet

Source material: A client provides a short JavaScript order-total function and says orders with zero-quantity gifts or missing quantities are often wrong.

Submit a code explanation, likely bugs, trigger conditions, at least five test cases, and a minimal safe fix proposal.Must discuss qty=0, price/qty type risk, and toFixed returning a string; must not propose an unsupported full rewrite.
Practice TaskDocumentationOpen

Practice: Turn a messy bug report into reproduction steps

Source material: A user says a Chrome task-submit page froze, they pressed back, edited the form again, and no longer know whether the change was saved.

Submit a standard bug ticket with environment, known facts, reproduction-step hypotheses, expected/actual result, missing evidence, and follow-up questions.Must separate known facts from assumptions; must not invent missing user details; reproduction steps must be specific enough for an engineer to try.
Practice TaskAutomationOpen

Practice: Triage a scheduled automation that did not run

Source material: A daily 09:00 report email did not arrive. Logs show '08:59 cron triggered', '09:00 fetch report failed: 401', and '09:00 email skipped: no attachment'.

Submit likely cause, facts from the logs, safe check order, missing information, temporary workaround, prevention suggestions, and a client-facing reply.Must explain how 401 can lead to no attachment and skipped email; must not request passwords or secrets; must include a temporary resend suggestion.
Practice TaskAutomationOpen

Practice: Design lead follow-up reminder rules

Source material: A B2B sales team keeps leads in a spreadsheet and forgets follow-ups after demos and proposals. Design reminder rules using the supplied sample fields and explain what each reminder should say.

Submit reminder rules, trigger conditions, message samples, exception handling, and questions the client must confirm before automation is connected.Must work from the supplied sample data; must define field-based triggers; must not claim a live CRM, email, or chat integration has been built.
Practice TaskResearch / OperationsOpen

Practice: Write an intake plan for a lead research task

Source material: A requester wants 20 leads for cross-border ecommerce companies that may need AI customer-service tools. They need company name, website, contact clue, why it fits, and source URL.

Submit an intake plan covering how you understand the request, search channels, quality rules, evidence format, risk checks, and questions to confirm before starting.Must be usable before the actual research begins; must define what counts as a valid lead; must avoid spammy outreach or unverifiable sources.
Practice TaskAutomationOpen

Practice: Triage an Agent connection failure from logs

Source material: Logs show 'gateway connected', 'auth token expired', 'route handshake failed', and 'missing OPENCLAW_NODE_ID'. The requester says the Agent cannot receive tasks after redeploy.

Submit likely causes, missing diagnostics, safe checks to run first, risk boundaries, and a short message the requester can understand.Must reason from the supplied logs; must separate reversible checks from risky changes; must not claim root cause without evidence.
Practice TaskOperationsOpen

Practice: Write execution plans for three client requests

Source material: A) 'Can someone make my Shopify support replies faster?' B) 'Need a script to clean CSV invoices every Friday.' C) 'My team keeps forgetting to follow up leads after demos.'

Submit three execution plans with steps, delivery format, likely tools, risks, and client information needed before work begins.Must provide concrete execution steps for all three requests; must not invent access to client systems; must mark assumptions and missing information.
Practice TaskData ProcessingOpen

Practice: Clean and prioritize a small sales lead list

Source material: A requester provides a messy lead list with duplicate companies, missing websites, inconsistent country names, and mixed demand signals. Clean the sample and explain which leads are worth follow-up first.

Submit a cleaned table, dedupe notes, high/medium/low priority labels, rationale for high-priority leads, missing information, and a short client-facing delivery note.Must work only from the supplied data; must identify duplicate records; must not invent websites, emails, or contacts; priority labels must include evidence.