Buckets:
| {"prompt":"make me a delivery routing dashboard","follow_up_prompts":["add drivers, stops, packages, ETAs, route map, failed deliveries, and proof of delivery","add rerouting when a driver falls behind or a package is missing","make the map and list stay in sync"]} | |
| {"prompt":"make me a map-based app for tracking community cleanup sites","follow_up_prompts":["add regions, pins, reports, photos, volunteer assignments, and status history","add duplicate report detection and moderation","make the map useful even when there are hundreds of pins"]} | |
| {"prompt":"make me an insurance claim tracker","follow_up_prompts":["add claim status, documents, messages, deadlines, payments, and denial reasons","add an appeal workflow with evidence collection","make the timeline and next steps really obvious"]} | |
| {"github_repo":"expressjs/express","prompt":"Investigate one small edge case in request handling, routing, middleware ordering, or response behavior. Write a failing test first if possible, then implement the minimal fix.","follow_up_prompts":["Simplify the patch while preserving the regression coverage.","Add a concise comment only if the behavior is non-obvious."]} | |
| {"prompt":"make me a dependency map for a software system","follow_up_prompts":["add services, databases, queues, owners, incidents, and dependency health","add impact analysis when one service goes down","make the graph readable instead of a hairball"]} | |
| {"prompt":"make me a notion-style notes app","follow_up_prompts":["add folders, tags, backlinks, and a nice search experience","add markdown shortcuts and a preview mode","make the editor feel polished and not janky"]} | |
| {"prompt":"make me a simple app builder","follow_up_prompts":["let me drag components onto a page and edit their properties","add save/load and generated code preview","make the editor feel less clunky and more like a real tool"]} | |
| {"prompt":"make me a study app for flashcards","follow_up_prompts":["add spaced repetition, decks, tags, and review stats","make the review flow fast and keyboard friendly","add import/export and handle duplicate cards nicely"]} | |
| {"prompt":"make me a dashboard for rss, youtube, and newsletters all in one place","follow_up_prompts":["add source management, read later, tags, and deduping","add AI summaries through an openai compatible endpoint","make the reading experience feel calm and distraction-free"]} | |
| {"prompt":"make me a simple github issues dashboard","follow_up_prompts":["add filtering by repo, label, assignee, stale status, and priority","add a triage mode for quickly categorizing issues","make it easy to plug into a real github api later"]} | |
| {"prompt":"make me an ai writing workspace","follow_up_prompts":["add documents, prompt presets, version history, and side-by-side outputs","make it support openai compatible endpoints and streaming","add a way to compare drafts and restore old versions"]} | |
| {"github_repo":"storybookjs/storybook","prompt":"Pick one small UI or developer-experience edge case in a Storybook example or addon area, implement a minimal improvement, and update tests or docs if practical.","follow_up_prompts":["Improve accessibility or keyboard behavior in the touched area.","Refactor the patch to reduce complexity."]} | |
| {"prompt":"Build a file upload API with file type validation, size limits, metadata storage, and safe error handling.","follow_up_prompts":["Harden the upload validation and security assumptions.","Add tests for invalid files, oversized files, and missing metadata."]} | |
| {"prompt":"Build a mobile-first recipe app with recipe cards, ingredients, cooking steps, favorites, and grocery list generation.","follow_up_prompts":["Improve the cooking mode UX.","Add better handling for dietary tags and missing ingredients."]} | |
| {"github_repo":"tastejs/todomvc","prompt":"Investigate one TodoMVC example and improve an edge case around input handling, empty todos, whitespace, filtering, or completion state. Add a focused regression test if the repo has a suitable test path.","follow_up_prompts":["Review the change for consistency across similar examples.","Improve the user-facing behavior without expanding the feature scope."]} | |
| {"prompt":"Build a weird but coherent web app for managing a community garden's shared tools, watering schedule, volunteer shifts, and harvest notes.","follow_up_prompts":["Make the UX friendlier for non-technical users.","Add edge cases around conflicting reservations and missed shifts."]} | |
| {"prompt":"Write a technical plan for replacing mock data in a frontend dashboard with a real API integration.","follow_up_prompts":["Add failure modes and fallback behavior.","Break the plan into small PR-sized steps."]} | |
| {"prompt":"Build a lightweight RAG prototype UI where users upload notes, search them semantically, and view cited snippets.","follow_up_prompts":["Improve the chunking/search explanation and citation display.","Add empty states and failure handling for bad uploads."]} | |
| {"prompt":"Build a SaaS landing page for an AI meeting-notes product with pricing, feature cards, social proof, FAQ, and a strong hero section.","follow_up_prompts":["Make the page feel more enterprise-ready.","Add better conversion-focused copy and visual hierarchy."]} | |
| {"prompt":"Build a small GraphQL API for a book review app with books, authors, reviews, ratings, and search.","follow_up_prompts":["Add pagination and improve resolver organization.","Add validation and tests for rating edge cases."]} | |
| {"prompt":"make a simple flappy bird type game"} | |
| {"prompt":"Build a word guessing game with daily challenge mode, keyboard input, hints, and shareable result text.","follow_up_prompts":["Improve the win/loss flow and feedback states.","Add local statistics and streak tracking."]} | |
| {"prompt":"Build a browser toy where users compose ambient music by placing objects on a canvas that each represent a loop or sound texture.","follow_up_prompts":["Improve the interaction design and visual feedback.","Add preset saving and better onboarding."]} | |
| {"github_repo":"vitejs/vite","prompt":"Explore Vite's repository and improve one small documentation, error-message, or test coverage gap around config loading, dev server behavior, plugins, or build output.","follow_up_prompts":["Add a focused regression test if the behavior is testable.","Make the developer-facing message clearer and more actionable."]} | |
| {"prompt":"Build a shopping cart and checkout mock flow with cart editing, shipping info, payment placeholder, order summary, and confirmation page.","follow_up_prompts":["Add validation and better checkout error states.","Improve trust signals and reduce checkout friction."]} | |
| {"prompt":"Build a fictional spaceship operations dashboard showing oxygen, fuel, crew status, maintenance alerts, and mission timeline.","follow_up_prompts":["Make the dashboard feel more cinematic but still usable.","Add alert prioritization and failure states."]} | |
| {"prompt":"Build a lightweight team OKR tracker with objectives, key results, confidence scores, owners, and weekly updates.","follow_up_prompts":["Add a better progress visualization system.","Improve the UX for reviewing stale or at-risk goals."]} | |
| {"github_repo":"fastapi/full-stack-fastapi-template","prompt":"Explore the full-stack template, identify one setup, configuration, or developer-experience rough edge, improve it, and update the relevant documentation.","follow_up_prompts":["Add a troubleshooting note for a likely local setup failure.","Make the change safer for someone cloning the template for a new project."]} | |
| {"prompt":"make me a local-first habit tracker","follow_up_prompts":["add streaks, weekly summaries, and gentle reminders","make it handle missed days in a forgiving way","add export/import so i don't lose my data"]} | |
| {"prompt":"Build a memory card matching game with multiple difficulty levels, timer, move counter, and win state.","follow_up_prompts":["Add animations and polish the game flow.","Improve accessibility and keyboard support."]} | |
| {"prompt":"make me a shift scheduling app","follow_up_prompts":["add employees, roles, availability, locations, shift swaps, overtime warnings, and approvals","add conflict detection and fairness indicators","make it easy to use from a phone"]} | |
| {"prompt":"Build a content calendar app for a solo creator with ideas, drafts, scheduled posts, channels, and performance notes.","follow_up_prompts":["Improve the planning workflow and calendar density.","Add reusable templates and better status tracking."]} | |
| {"prompt":"make me an accommodations tracker for a school","follow_up_prompts":["add students, accommodation plans, teacher acknowledgements, deadlines, and review meetings","add audit history for who changed what","make sensitive information visible only to the right roles"]} | |
| {"prompt":"Build a small tower defense prototype where enemies follow a path and the player places towers to stop them.","follow_up_prompts":["Add upgrade mechanics and better balancing.","Refactor the game logic so it is easier to extend."]} | |
| {"prompt":"make me a warehouse picking app","follow_up_prompts":["add pick lists, bin locations, barcode scanning placeholders, substitutions, and packing status","add conflict handling when inventory counts don't match","make it fast for workers using handheld devices"]} | |
| {"prompt":"make me a lab sample tracking app","follow_up_prompts":["add samples, freezer locations, chain of custody, tests, results, and disposal status","add barcode placeholders and audit history","make it hard to accidentally lose or overwrite sample data"]} | |
| {"prompt":"Write a compact implementation plan for adding role-based access control to an existing SaaS dashboard.","follow_up_prompts":["Revise it to minimize migration risk.","Add test cases and rollout checks."]} | |
| {"prompt":"make me an audit evidence collection dashboard","follow_up_prompts":["add controls, evidence requests, owners, due dates, reviewers, and status history","add reminders and partial submission handling","make it clear what evidence is missing or stale"]} | |
| {"prompt":"make me an inventory audit app","follow_up_prompts":["add locations, item counts, discrepancies, photos, approvals, and export","add barcode scanning placeholders and offline collection","make reconciliation easy when two people count different numbers"]} | |
| {"prompt":"make me a parent teacher communication portal","follow_up_prompts":["add announcements, messages, conferences, translations, and read receipts","add privacy controls so parents only see the right student info","make it work for multilingual families"]} | |
| {"prompt":"make me an app to manage a small community garden","follow_up_prompts":["add plots, members, watering schedules, shared tools, harvest logs, and volunteer shifts","add conflict handling for tools and missed responsibilities","make it usable for older non-technical members"]} | |
| {"prompt":"make me a community resource directory","follow_up_prompts":["add organizations, services, eligibility, locations, languages, hours, and update status","add stale listing detection and suggested edits","make search work for people who don't know the exact service name"]} | |
| {"prompt":"make me a pet care app","follow_up_prompts":["add feeding, medication, vet visits, grooming, reminders, and multiple pets","add emergency info and shared access for family members","make it cute but still clear when something is urgent"]} | |
| {"prompt":"Build a lightweight CMS editor for blog posts with title, slug, rich text placeholder, preview, draft/publish status, and revision history mock.","follow_up_prompts":["Improve the editing flow and unsaved changes handling.","Add validation for slugs and publish readiness."]} | |
| {"prompt":"make me a tetris-style game"} | |
| {"prompt":"make me a dataset labeling tool","follow_up_prompts":["add labeling queues, keyboard shortcuts, review mode, disagreements, and export","add quality checks for inconsistent labels","make the workflow fast for hundreds of examples"]} | |
| {"github_repo":"microsoft/playwright","prompt":"Investigate one small test-runner edge case or confusing error path, add focused coverage if practical, and implement a narrowly scoped improvement.","follow_up_prompts":["Refactor the fix to match surrounding test-runner conventions.","Improve the error message so it suggests the next action."]} | |
| {"prompt":"Build an email triage mock app that groups messages by urgency, sender, topic, and required action.","follow_up_prompts":["Improve the UX for reviewing and correcting classifications.","Add safe fallback behavior for ambiguous messages."]} | |
| {"prompt":"make me a clinic appointment dashboard","follow_up_prompts":["add check-in status, provider schedule, wait time estimates, and room assignment","add conflict handling when appointments run late or rooms change","make it usable on a tablet at the front desk"]} | |
| {"prompt":"make me a recipe website for my family","follow_up_prompts":["add categories, stories, photos, print mode, and favorites","add a way for relatives to submit recipes","make it feel personal and not like a generic food blog"]} | |
| {"prompt":"Build a CRM-style web app for tracking leads, deals, contacts, and follow-up tasks.","follow_up_prompts":["Add role-based views for sales reps and managers.","Make the interaction design feel closer to a real production app."]} | |
| {"prompt":"make me a kanban board for content creation","follow_up_prompts":["add ideas, drafts, scheduled posts, platforms, and performance notes","add a calendar view and reusable content templates","make it help me spot what content is stuck"]} | |
| {"prompt":"make me a customer support dashboard","follow_up_prompts":["add ticket priority, statuses, tags, canned replies, and assignment","add a triage view for urgent tickets","make it obvious what changed when a ticket updates"]} | |
| {"prompt":"make me a project management app like trello but cleaner","follow_up_prompts":["add drag and drop, labels, due dates, and assigned people","add a timeline/calendar view without making the app messy","make it work nicely on mobile too"]} | |
| {"prompt":"Build an authentication service mock with registration, login, password reset, sessions, and protected routes.","follow_up_prompts":["Review the implementation for likely security issues.","Improve tests around expired sessions and invalid tokens."]} | |
| {"prompt":"Build a browser-based puzzle game where players rotate pipe tiles to connect water from a start tile to an end tile.","follow_up_prompts":["Improve the game feel with animations and clearer feedback.","Add levels, scoring, and local progress saving."]} | |
| {"github_repo":"pallets/click","prompt":"Investigate an edge case involving options, arguments, command groups, help text, or error messages. Reproduce it with a test and implement the smallest reasonable fix.","follow_up_prompts":["Refactor the fix to fit the surrounding parser design.","Explain why the behavior is consistent with Click's CLI model."]} | |
| {"prompt":"make me a car maintenance tracker","follow_up_prompts":["add vehicles, mileage, service history, reminders, costs, and documents","add upcoming maintenance predictions based on mileage","make it easy to add a service record from my phone"]} | |
| {"prompt":"Debug this scenario: an API endpoint occasionally creates duplicate records when users submit the same form twice. Identify the likely race/idempotency issue and fix it.","follow_up_prompts":["Add tests for duplicate submissions.","Improve the API response for repeated requests."]} | |
| {"prompt":"Build a product catalog page for an online clothing store with filters, sorting, product cards, wishlist, and responsive layout.","follow_up_prompts":["Improve the filtering UX and empty result states.","Make the product cards feel more premium and useful."]} | |
| {"prompt":"make me a language learning app","follow_up_prompts":["add lessons, vocabulary, quizzes, progress, and streaks","add speaking/listening placeholders and better review flow","make it feel playful without being childish"]} | |
| {"prompt":"make me a donation tracking dashboard","follow_up_prompts":["add donors, campaigns, pledges, recurring gifts, receipts, and thank-you tasks","add deduping for messy donor imports","make the reporting useful without feeling salesy"]} | |
| {"prompt":"make me a volunteer coordination app for a nonprofit","follow_up_prompts":["add shifts, skills, availability, background check status, reminders, and check-in","add coordinator and volunteer views","make no-shows and last-minute replacements easier to handle"]} | |
| {"prompt":"Build a landing page for a local gym with class schedules, trainer profiles, membership tiers, and a free trial CTA.","follow_up_prompts":["Make the layout more energetic and conversion-focused.","Add accessibility improvements and better responsive behavior."]} | |
| {"github_repo":"eslint/eslint","prompt":"Explore ESLint's repository and improve one small rule documentation, test case, error message, or developer-experience issue around configuration or lint output.","follow_up_prompts":["Add a focused test for the exact linting behavior.","Make the docs clearer without changing rule semantics."]} | |
| {"prompt":"make me a portfolio website, im a graphic designer","follow_up_prompts":["make it look colorful, more unique, and less AI generated","add a project case study layout that feels professional","make the mobile version feel just as intentional as desktop"]} | |
| {"prompt":"make me a habit app for a group of friends","follow_up_prompts":["add shared challenges, streaks, reactions, and weekly leaderboards","make privacy controls so not everything is public","make it feel encouraging instead of competitive and annoying"]} | |
| {"prompt":"make me an app for people with memory issues to manage daily routines","follow_up_prompts":["add simple checklists, gentle reminders, caregiver view, and confirmation history","make the UI very calm and hard to misclick","add multilingual and large-text options"]} | |
| {"prompt":"make me a wedding planning app","follow_up_prompts":["add guests, RSVP tracking, budget, vendors, seating, and timeline","add conflict warnings for seating and missing RSVPs","make it pretty but still useful under stress"]} | |
| {"prompt":"make me a little snake game"} | |
| {"prompt":"Build a mobile-first habit tracker with daily check-ins, streaks, reminders mock, and weekly progress summary.","follow_up_prompts":["Improve the one-handed mobile UX.","Add better progress feedback and missed-day recovery."]} | |
| {"prompt":"make me a tower defense game","follow_up_prompts":["add different towers, enemy waves, upgrades, and pathing","make it balanced enough to be fun for a few levels","refactor the game logic so adding new enemies is easy"]} | |
| {"prompt":"make me a document search app","follow_up_prompts":["add upload, chunking, search results, filters, and source previews","wire in embeddings through an openai compatible endpoint","make the citations and source snippets really clear"]} | |
| {"prompt":"Write a short project plan for building a solo-developer MVP of a habit tracking app, including milestones, risks, and testing priorities.","follow_up_prompts":["Revise it for a two-week timeline.","Add a risk checklist and cut-scope recommendations."]} | |
| {"github_repo":"microsoft/playwright","prompt":"Explore Playwright's docs or examples and improve one small developer-experience issue around test setup, selectors, tracing, retries, or debugging.","follow_up_prompts":["Add a minimal example that demonstrates the improved guidance.","Make the troubleshooting path clearer for a new user."]} | |
| {"prompt":"make me a PDF intake tool for forms people email in","follow_up_prompts":["add upload, OCR text preview, extracted fields, confidence warnings, and manual correction","add reviewer approval before data is imported","make the audit history clear for every extracted field"]} | |
| {"prompt":"make me a classroom behavior support app","follow_up_prompts":["add incident notes, positive observations, interventions, parent contact, and trend charts","add safeguards so it doesn't become punitive or biased","make the reports useful for support meetings"]} | |
| {"prompt":"make me a browser game where you run a coffee shop","follow_up_prompts":["add customers, recipes, inventory, upgrades, and daily profit","make the simulation more interesting without making it too complicated","add a tutorial and better feedback when things go wrong"]} | |
| {"prompt":"Debug this scenario: a background job retry system retries forever for some permanent failures. Fix the retry policy and add clear dead-letter behavior.","follow_up_prompts":["Add tests for retry limits and permanent errors.","Improve logging and observability for failed jobs."]} | |
| {"prompt":"make me a lesson planning app for teachers","follow_up_prompts":["add units, standards, activities, materials, assessments, and differentiation notes","add a curriculum map view across the semester","make it easy to reuse and adapt old lessons"]} | |
| {"github_repo":"gothinkster/node-express-realworld-example-app","prompt":"Trace one feature path in the RealWorld API, such as authentication, article creation, comments, or favorites. Make a small maintainability improvement without changing the public API.","follow_up_prompts":["Add focused tests around the refactored path.","Document the feature flow for a new contributor."]} | |
| {"prompt":"Build a small code generation CLI that scaffolds a React component with tests, styles, and story/demo file.","follow_up_prompts":["Add safer file overwrite behavior.","Improve the generated code quality and naming conventions."]} | |
| {"prompt":"make me a home care coordination app","follow_up_prompts":["add care plans, visit notes, tasks, medication reminders, and family updates","add permissions for caregivers, family, and clinicians","make it work decently when someone fills it out from a phone in the field"]} | |
| {"prompt":"make me a performance review workflow app","follow_up_prompts":["add self reviews, manager reviews, peer feedback, calibration, goals, and acknowledgements","add privacy controls and edit history","make the writing prompts less corporate and more useful"]} | |
| {"prompt":"make me a browser-based city builder game","follow_up_prompts":["add resources, buildings, upgrades, population, and disasters","make the simulation tick system clean and not spaghetti","add saving/loading and a better tutorial"]} | |
| {"github_repo":"vitejs/vite","prompt":"Investigate one small edge case involving path resolution, environment variables, plugin ordering, or dev-server errors. Reproduce it with a test and implement a minimal fix.","follow_up_prompts":["Review the fix for cross-platform path behavior.","Add documentation only if the behavior is user-visible."]} | |
| {"prompt":"Build a report generator app that takes structured sales data and outputs a concise executive summary with charts.","follow_up_prompts":["Add better handling for missing or suspicious data.","Improve the report layout and export experience."]} | |
| {"prompt":"make me an apartment hunting tracker","follow_up_prompts":["add listings, prices, neighborhoods, notes, pros/cons, tours, and application status","add comparison views and reminders for follow-ups","make it help me choose instead of just storing links"]} | |
| {"prompt":"make me a website for a fake design agency","follow_up_prompts":["make it less template-looking and more bold","add case studies that feel realistic","make the animations tasteful instead of overdone"]} | |
| {"prompt":"can you build me a 2048 game but make it look nice"} | |
| {"prompt":"make me an image review and moderation queue","follow_up_prompts":["add image metadata, tags, reviewer decisions, appeals, and audit history","add blur-by-default for sensitive content","make keyboard review fast but hard to accidentally approve things"]} | |
| {"prompt":"make me a classroom dashboard for teachers","follow_up_prompts":["add students, assignments, grades, attendance, and notes","add a view for spotting students who might need help","make it simple enough for a tired teacher to use quickly"]} | |
| {"prompt":"Debug this scenario: a browser game increments the score twice when two events happen close together. Find and fix the state-management bug.","follow_up_prompts":["Add a regression test or deterministic reproduction if possible.","Refactor the scoring logic to make duplicate increments harder."]} | |
| {"prompt":"make me a reproducibility checklist app","follow_up_prompts":["add datasets, code versions, environments, parameters, results, and reviewer signoff","add warnings for missing artifacts or undocumented steps","make the final report easy to export"]} | |
| {"prompt":"make me a pokemon-style battle simulator","follow_up_prompts":["add moves, types, health, turn order, and status effects","add a team builder and simple enemy AI","make the battle log clear and fun to read"]} | |
| {"github_repo":"fastapi/full-stack-fastapi-template","prompt":"Inspect the backend portion of this template and improve one small validation, test, or error-handling path while preserving the template's architecture.","follow_up_prompts":["Add a regression test for the changed behavior.","Improve the implementation so it remains clear for template users."]} | |
| {"prompt":"make me an incident reporting system","follow_up_prompts":["add intake forms, severity, attachments, witnesses, investigation notes, and corrective actions","add role-based visibility and escalation workflows","make the audit trail tamper-evident in the UI"]} | |
| {"prompt":"Debug this scenario: a React dashboard sometimes shows stale data after switching filters quickly. Investigate likely causes and implement a robust fix.","follow_up_prompts":["Add tests that reproduce the stale data issue.","Improve loading and cancellation behavior in the UI."]} | |
| {"prompt":"make me a rental application review app","follow_up_prompts":["add applicants, documents, screening status, notes, approvals, and rejection reasons","add fair housing-safe review guidance and audit logs","make it easy to compare applications without exposing too much PII"]} | |
| {"prompt":"make me a simple discord clone","follow_up_prompts":["add servers, channels, direct messages, and online indicators","make it feel real with message states, typing indicators, and empty states","clean up the architecture so realtime support could be added later"]} | |
| {"prompt":"make me a simple internal wiki","follow_up_prompts":["add pages, folders, search, tags, links between pages, and edit history","add permissions for public, team, and private pages","make it easy to navigate when there are lots of pages"]} | |
| {"prompt":"make me a privacy request management dashboard","follow_up_prompts":["add requests, identity verification status, deadlines, data sources, approvals, and deletion/export logs","add role-based permissions and legal hold exceptions","make the workflow understandable for support agents"]} | |
| {"prompt":"Build a background job queue simulation with job creation, retries, dead-letter handling, and a dashboard endpoint.","follow_up_prompts":["Improve observability and failure handling.","Add tests for retry limits and duplicate jobs."]} | |
| {"prompt":"make me a medication reminder app for caregivers","follow_up_prompts":["add multiple patients, medication schedules, refill reminders, and missed dose tracking","add a caregiver handoff log so family members know what happened","make the alerts clear without making the app stressful"]} | |
| {"prompt":"make me a landing page for an ai startup","follow_up_prompts":["make it sound less generic and more credible","add pricing, FAQ, testimonials, and product screenshots","make the visual style feel fresh without using the usual blue gradient AI look"]} | |
| {"github_repo":"chalk/chalk","prompt":"Investigate how Chalk handles nested styles or unsupported color levels. Make a small correctness, performance, or documentation improvement and update the relevant tests.","follow_up_prompts":["Review the change for backwards compatibility.","Add a concise explanation for the edge case."]} | |
| {"prompt":"Debug this scenario: a checkout form accepts invalid postal codes and sometimes shows two different error messages for the same field. Fix the validation flow.","follow_up_prompts":["Refactor the form validation to be easier to maintain.","Improve the user-facing error messages."]} | |
| {"github_repo":"prettier/prettier","prompt":"Explore Prettier's codebase and pick one small documentation, fixture, or formatting edge case improvement. Keep the patch extremely focused.","follow_up_prompts":["Add or update the relevant formatting snapshot/fixture.","Explain why the output is more consistent with nearby cases."]} | |
| {"prompt":"make me a recipe app","follow_up_prompts":["add meal planning, grocery lists, and pantry tracking","make it suggest recipes based on what i already have","make the cooking mode really nice on a phone in the kitchen"]} | |
| {"prompt":"Build a React dashboard for a subscription analytics SaaS showing MRR, churn, active users, revenue by plan, and recent customer activity.","follow_up_prompts":["Add loading, empty, and error states throughout the dashboard.","Improve the data visualization layout and information hierarchy."]} | |
| {"prompt":"make me a dashboard for monitoring servers","follow_up_prompts":["add uptime, CPU, memory, disk, logs, alerts, and incident history","make alert states impossible to miss but not ugly","add mock realtime updates and make the app not fall apart with lots of servers"]} | |
| {"prompt":"make me an accessibility-first todo app for screen reader users","follow_up_prompts":["make every action keyboard accessible and clearly announced","add low-vision mode and reduced cognitive load settings","test the app's empty, error, and reorder states for accessibility"]} | |
| {"prompt":"make me a bug tracker","follow_up_prompts":["add projects, issues, labels, priority, assignees, comments, and status changes","add a useful dashboard for what needs attention","make it good for both developers and non-technical users"]} | |
| {"github_repo":"hoppscotch/hoppscotch","prompt":"Explore Hoppscotch as an API tooling app, identify one small UX improvement around request history, collections, response display, auth, or error states, implement it, and update tests if practical.","follow_up_prompts":["Make the improved flow more accessible and keyboard-friendly.","Add a focused test or story/example for the changed UI."]} | |
| {"prompt":"make me an app for tracking job applications","follow_up_prompts":["add company stages, contacts, reminders, notes, and interview prep","add a dashboard that shows what needs follow up this week","make it easy to update without feeling like data entry hell"]} | |
| {"prompt":"Build a customer support inbox app with ticket list, ticket detail pane, status changes, tags, priority indicators, and canned replies.","follow_up_prompts":["Improve the workflow for triaging urgent tickets.","Add empty states, error states, and optimistic UI where appropriate."]} | |
| {"prompt":"make me a low-vision friendly calendar app","follow_up_prompts":["add large text, high contrast, keyboard navigation, and clear focus states","add voice-friendly labels and simple day/week views","make dense schedules easier to understand without tiny text"]} | |
| {"prompt":"make me a grant application tracker","follow_up_prompts":["add funders, deadlines, requirements, documents, budgets, reviewers, and submission status","add reminders and risk flags for missing materials","make it work for a small nonprofit team"]} | |
| {"prompt":"Build a migration helper CLI that renames files and updates imports according to a config file.","follow_up_prompts":["Add dry-run output and rollback guidance.","Add tests for nested directories and ambiguous imports."]} | |
| {"prompt":"Build a conference event website with speakers, agenda, sponsor tiers, venue details, and registration CTA.","follow_up_prompts":["Add a polished schedule view and improve information density.","Make the design feel more professional for a tech audience."]} | |
| {"github_repo":"pallets/click","prompt":"Explore the Click codebase, identify one small CLI behavior, documentation, or test coverage improvement, implement it, and keep the patch narrowly scoped.","follow_up_prompts":["Add a focused test using Click's testing utilities.","Improve the docs or changelog-style note for the behavior."]} | |
| {"prompt":"make me an energy usage dashboard","follow_up_prompts":["add time-series charts, meters, alerts, building zones, and unusual usage detection","add comparison between days, weeks, and locations","make spikes and missing sensor data easy to understand"]} | |
| {"github_repo":"storybookjs/storybook","prompt":"Explore Storybook's repository and identify one small documentation or example improvement that helps users understand stories, controls, args, addons, or testing.","follow_up_prompts":["Make the example more realistic while keeping it minimal.","Add a short troubleshooting note for a common mistake."]} | |
| {"prompt":"Build a marketplace seller dashboard with listings, messages, orders, payouts, and performance metrics.","follow_up_prompts":["Improve the seller workflow for urgent actions.","Add better table filtering and mobile behavior."]} | |
| {"prompt":"make me an experiment logbook","follow_up_prompts":["add protocols, runs, parameters, observations, attachments, results, and reproducibility notes","add versioning so old protocol details are preserved","make it useful for reviewing failed experiments"]} | |
| {"prompt":"Build a simple annotation tool for labeling text snippets as positive, negative, neutral, or uncertain.","follow_up_prompts":["Add keyboard shortcuts and review mode.","Improve the data export format and validation."]} | |
| {"github_repo":"gothinkster/node-express-realworld-example-app","prompt":"Explore this RealWorld Node/Express API example, identify one small validation or error-handling gap, patch it, and update the tests or API behavior notes.","follow_up_prompts":["Refactor the validation logic to reduce duplication.","Add a regression test for the most important edge case."]} | |
| {"prompt":"make me a fake robinhood-style trading app","follow_up_prompts":["add watchlists, portfolio performance, and stock detail pages with charts","make the data layer easy to swap out later for a real API","add warnings and safe UX so it doesn't feel like real financial advice"]} | |
| {"prompt":"Build a community forum app with categories, threads, replies, moderation states, and user badges.","follow_up_prompts":["Improve moderation workflows and edge cases.","Add better empty states and thread sorting."]} | |
| {"prompt":"Build a markdown link checker CLI that scans files, detects broken internal anchors, and reports invalid links.","follow_up_prompts":["Add ignore patterns and better reporting.","Improve tests around relative paths and duplicate headings."]} | |
| {"github_repo":"refinedev/refine","prompt":"Explore refine as a React framework for internal tools and dashboards. Pick one small documentation, example, or developer-experience improvement related to resources, forms, tables, or routing.","follow_up_prompts":["Add a concise example showing the improved behavior.","Make the wording clearer for developers building admin panels."]} | |
| {"prompt":"make me a school field trip planner","follow_up_prompts":["add students, permission slips, emergency contacts, chaperones, buses, payments, and checklists","add day-of check-in and missing student alerts","make it usable offline during the trip"]} | |
| {"prompt":"build me a minesweeper clone"} | |
| {"github_repo":"tastejs/todomvc","prompt":"Explore the React example in this repository and modernize one small piece of implementation or documentation without changing the app's expected behavior.","follow_up_prompts":["Add or update a minimal test/example for the changed behavior.","Make the change easier for a new contributor to understand."]} | |
| {"prompt":"Build a product launch page for a smart home device with hero section, feature breakdown, specs, comparison section, and preorder CTA.","follow_up_prompts":["Make the product storytelling stronger.","Add trust signals and improve the purchase flow."]} | |
| {"prompt":"Build a tiny app for tracking a tabletop RPG campaign with characters, quests, locations, inventory, and session notes.","follow_up_prompts":["Improve the information architecture for long campaigns.","Add search, tags, and relationship links."]} | |
| {"prompt":"Build a browser automation planning UI where users describe a repetitive web task and the app shows a step-by-step plan before execution.","follow_up_prompts":["Add confirmation checkpoints for risky actions.","Improve error recovery and plan editing UX."]} | |
| {"prompt":"make me a mental wellness check-in app","follow_up_prompts":["add mood tracking, habits, journal prompts, crisis resources, and trend summaries","make the wording supportive without pretending to be a therapist","add privacy-first export and delete controls"]} | |
| {"prompt":"make me a vendor compliance tracker","follow_up_prompts":["add vendors, questionnaires, documents, risk ratings, renewals, and approvals","add stale evidence and missing document warnings","make it easy to compare vendor risk over time"]} | |
| {"github_repo":"chalk/chalk","prompt":"Explore Chalk's codebase and tests, identify one small edge case or documentation gap around color support, template literals, terminal detection, or styling behavior, and improve it.","follow_up_prompts":["Add a focused test or example for the changed behavior.","Simplify the implementation while preserving compatibility."]} | |
| {"prompt":"Build a calendar scheduling assistant UI that suggests meeting slots based on availability, duration, timezone, and attendee constraints.","follow_up_prompts":["Add better conflict explanations and timezone clarity.","Improve the rescheduling flow."]} | |
| {"prompt":"make me a workflow automation builder","follow_up_prompts":["add triggers, actions, conditions, and a visual flow editor","add run history and clear errors when a step fails","make the internal data model clean enough to add more integrations later"]} | |
| {"prompt":"Build a kanban project management app with boards, columns, draggable cards, labels, due dates, and assignees.","follow_up_prompts":["Add keyboard-accessible interactions where reasonable.","Add a compact activity log and improve the card detail view."]} | |
| {"prompt":"make me a lease tracking dashboard","follow_up_prompts":["add units, tenants, lease terms, renewals, deposits, documents, and important deadlines","add alerts for expiring leases and missing documents","make sensitive tenant data role-restricted"]} | |
| {"prompt":"Debug this scenario: a CLI works on simple file paths but fails on paths with spaces, unicode characters, or nested directories. Fix the issue and add tests.","follow_up_prompts":["Improve the CLI output for failed files.","Add a dry-run mode to preview changes safely."]} | |
| {"prompt":"Build a beautiful one-page website for a boutique interior design studio with services, project highlights, testimonials, and contact form.","follow_up_prompts":["Make the design feel more luxurious and minimal.","Improve the project showcase section."]} | |
| {"prompt":"Build a REST API for a habit tracking app with users, habits, daily completions, streak calculation, and basic validation.","follow_up_prompts":["Add tests for edge cases around streaks and missed days.","Improve the API error format and input validation."]} | |
| {"prompt":"make me a multiplayer drawing game","follow_up_prompts":["add rooms, timers, guessing, scores, and player names","make the drawing feel smooth and handle disconnects gracefully","add a lobby flow so friends can join easily"]} | |
| {"prompt":"Build an idle clicker game about running a tiny coffee shop with upgrades, passive income, and milestone unlocks.","follow_up_prompts":["Make the progression curve more satisfying.","Improve the UI so upgrades and milestones are clearer."]} | |
| {"prompt":"make me an offline inspection app","follow_up_prompts":["add forms, photos, GPS location, drafts, sync status, and conflict resolution","make it obvious what has synced and what hasn't","add a safe way to recover from partial saves"]} | |
| {"prompt":"Build an app for planning a neighborhood garage sale with households, item categories, map zones, volunteer shifts, and printable signs.","follow_up_prompts":["Improve the workflow for organizers.","Add export/print-friendly views."]} | |
| {"prompt":"make me a tiny ecommerce store","follow_up_prompts":["add cart, checkout flow, product filters, and order confirmation","add admin screens for products, inventory, and orders","make the checkout feel trustworthy and handle errors well"]} | |
| {"prompt":"Build a CLI that scans a project for TODO comments and outputs a grouped report by file, tag, and priority.","follow_up_prompts":["Add JSON output and filtering options.","Refactor the parser and add tests for edge cases."]} | |
| {"prompt":"Build a small chess tactics trainer UI with a board, move input, puzzle success/failure feedback, and puzzle list.","follow_up_prompts":["Add a cleaner puzzle review experience.","Improve the board interactions and invalid-move handling."]} | |
| {"prompt":"make me a machine telemetry dashboard","follow_up_prompts":["add live sensor readings, status history, alerts, maintenance notes, and device filters","add streaming updates and offline device handling","make the time-series views not fall apart with lots of data"]} | |
| {"prompt":"make me a prompt library app","follow_up_prompts":["add categories, tags, favorites, variables, version history, and search","add a way to test prompts against different models","make it support openai compatible endpoints"]} | |
| {"prompt":"make me a symptom journal app","follow_up_prompts":["add charts, triggers, medication notes, appointments, and export for a doctor","add privacy controls and a way to redact sensitive notes before sharing","make the data entry fast when someone feels bad"]} | |
| {"prompt":"make me a dashboard for my small business","follow_up_prompts":["add revenue, expenses, tasks, customers, inventory, and upcoming deadlines","make it show what needs attention today","make the visual style friendlier and less corporate"]} | |
| {"prompt":"Build a mobile-first personal finance app screen with balances, recent transactions, budget progress, and spending alerts.","follow_up_prompts":["Improve the visual hierarchy for small screens.","Add empty, loading, and error states."]} | |
| {"prompt":"make me a grading workflow app","follow_up_prompts":["add rubrics, submissions, comments, late policy, regrade requests, and export","add a reviewer mode for teaching assistants","make bulk grading fast without losing feedback quality"]} | |
| {"prompt":"make me a web dashboard that aggregates rss feeds and presents them really nicely in one dashboard","follow_up_prompts":["wire in a summarization engine that summarizes all incoming articles via openai compatible endpoint so we only see summaries of articles in the dashboard","add categories, saved articles, and a way to hide sources i don't like","make it feel more like a polished news command center and less like a basic list"]} | |
| {"prompt":"make me a music practice tracker","follow_up_prompts":["add instruments, practice sessions, goals, exercises, and progress charts","add a weekly review that tells me what to work on","make logging practice take almost no effort"]} | |
| {"prompt":"Build a warm, modern landing page for a neighborhood bakery with featured products, opening hours, testimonials, and a call-to-action for custom orders.","follow_up_prompts":["Make the theming feel warmer and more handcrafted.","Improve the UI/UX for first-time visitors deciding what to order."]} | |
| {"prompt":"make me an internal training portal","follow_up_prompts":["add courses, lessons, quizzes, assignments, due dates, certificates, and manager reports","add accommodations and multilingual content support","make progress tracking clear without being creepy"]} | |
| {"prompt":"make me an app that helps me organize a hackathon","follow_up_prompts":["add teams, projects, judging, schedule, mentors, announcements, and submissions","add admin views for organizers and public views for participants","make the judging flow hard to mess up"]} | |
| {"github_repo":"hoppscotch/hoppscotch","prompt":"Trace one API request-building flow in Hoppscotch and improve a small edge case involving empty input, invalid headers, copied output, or confusing state transitions.","follow_up_prompts":["Improve the user feedback for the edge case.","Refactor the touched code to better match the surrounding architecture."]} | |
| {"prompt":"make me a little app to manage my whole life","follow_up_prompts":["split it into tasks, notes, calendar stuff, and habits without making it overwhelming","add a daily dashboard that tells me what to focus on","make it feel calmer and less like enterprise software"]} | |
| {"prompt":"Build a command-line tool that organizes a downloads folder by file type, date, and duplicate detection.","follow_up_prompts":["Add a dry-run mode and clearer CLI output.","Improve error handling and add tests for tricky filenames."]} | |
| {"prompt":"Build an AI prompt playground UI with prompt history, model settings, output comparison, and saved presets.","follow_up_prompts":["Improve the workflow for comparing multiple runs.","Add better organization for saved prompts."]} | |
| {"prompt":"make me a workout tracker","follow_up_prompts":["add custom routines, workout history, progress charts, and rest timers","make it handle supersets and weird exercise names","make the UI fast to use while i'm actually at the gym"]} | |
| {"prompt":"Build a physics sandbox toy where users can spawn bouncing balls, adjust gravity, clear the canvas, and save a preset.","follow_up_prompts":["Improve performance when many objects are on screen.","Add a more polished control panel and better defaults."]} | |
| {"prompt":"Build a small local dev server CLI that serves static files, watches changes, and prints useful request logs.","follow_up_prompts":["Add better handling for missing files and port conflicts.","Improve the developer experience of the command output."]} | |
| {"prompt":"Build a CLI log analyzer that reads server logs, detects common error patterns, and prints a compact summary.","follow_up_prompts":["Add support for multiple log formats.","Improve performance for large files."]} | |
| {"prompt":"Build an admin dashboard for a small e-commerce store with orders, inventory alerts, sales summary, and customer support tickets.","follow_up_prompts":["Add filters and sorting for the main tables.","Improve the UX for handling urgent orders and low-stock products."]} | |
| {"prompt":"Build a login and onboarding flow for a B2B SaaS app with sign up, email verification mock, workspace creation, and invite teammates screen.","follow_up_prompts":["Harden the form validation and error handling.","Make the onboarding feel more guided and less overwhelming."]} | |
| {"prompt":"Build a GitHub issue triage dashboard that groups issues by bug, feature, docs, priority, and stale status.","follow_up_prompts":["Add bulk actions and better filtering.","Improve the logic for detecting stale or duplicate issues."]} | |
| {"prompt":"build me a little ai chat app","follow_up_prompts":["make it support openai compatible endpoints and model selection","add chat history, system prompts, and streaming responses","make errors from the API actually understandable to normal users"]} | |
| {"prompt":"make me an audio transcript review tool","follow_up_prompts":["add audio playback, transcript editing, speaker labels, timestamps, confidence scores, and export","add keyboard shortcuts and change history","make it handle messy transcripts with missing speakers"]} | |
| {"prompt":"make me a multilingual content approval app","follow_up_prompts":["add source text, translations, locale status, reviewers, comments, and publish readiness","add RTL layout checks and regional formatting warnings","make missing translations and stale translations obvious"]} | |
| {"prompt":"make me a field service dispatch app","follow_up_prompts":["add technicians, jobs, routes, parts, customer notes, photos, and completion signatures","add offline mode and sync conflict handling","make it clear what a tech needs to do next on-site"]} | |
| {"prompt":"make me a web app that turns CSV files into nice dashboards","follow_up_prompts":["add upload, column detection, charts, filters, and missing value warnings","make it handle ugly real-world CSVs without crashing","add a report export view that looks professional"]} | |
| {"prompt":"make me a case management app for a community nonprofit","follow_up_prompts":["add clients, needs, referrals, notes, documents, consent, and follow-ups","add strict permissions and redaction for sensitive case notes","make the timeline of help provided easy to review"]} | |
| {"prompt":"Build a review and ratings component for an e-commerce site with sorting, filtering, helpful votes, and review summary stats.","follow_up_prompts":["Add edge cases for no reviews and suspicious review distribution.","Improve accessibility and information hierarchy."]} | |
| {"prompt":"Build a rate limiter middleware with configurable limits, per-user and per-IP modes, and clear error responses.","follow_up_prompts":["Add tests for boundary conditions and reset behavior.","Improve the design so storage backends can be swapped later."]} | |
| {"prompt":"make me a tenant maintenance request app","follow_up_prompts":["add tenants, properties, requests, photos, priorities, vendors, appointments, and completion notes","add landlord, tenant, and vendor views","make status updates and delays less confusing"]} | |
| {"prompt":"make me a job site safety checklist app","follow_up_prompts":["add inspections, hazards, photos, corrective actions, signoffs, and incident reports","add supervisor review and audit trails","make it work well on a phone with bad connectivity"]} | |
| {"github_repo":"prettier/prettier","prompt":"Investigate a narrow formatting edge case in one supported language, add a fixture that captures the behavior, and make the smallest reasonable improvement.","follow_up_prompts":["Reduce the patch size while preserving the fixture.","Update docs only if the behavior is user-visible."]} | |
| {"prompt":"make me a form builder","follow_up_prompts":["add field types, validation rules, preview mode, and submissions table","add conditional logic between fields","make it easy to export the schema and responses"]} | |
| {"prompt":"make me a tutoring marketplace","follow_up_prompts":["add tutor profiles, subjects, availability, bookings, payments mock, reviews, and messaging","add trust and safety flows for reports and cancellations","make search and matching actually useful"]} | |
| {"prompt":"Build a restaurant website for a small ramen shop with menu sections, location info, hours, photo-forward design, and online reservation call-to-action.","follow_up_prompts":["Improve the mobile ordering/reservation flow.","Make the design feel more authentic and less generic."]} | |
| {"prompt":"Build a mobile-first travel planner with itinerary days, places to visit, budget notes, packing checklist, and offline-friendly layout.","follow_up_prompts":["Improve the itinerary editing experience.","Add conflict warnings and better day overview."]} | |
| {"prompt":"make me a PTO planning app for a company","follow_up_prompts":["add balances, requests, approvals, team calendar, blackout dates, and coverage warnings","add manager and employee views","make conflict resolution fair and easy to understand"]} | |
| {"prompt":"make me a mutual aid request board","follow_up_prompts":["add requests, offers, location areas, urgency, matching, status, and moderation","add privacy controls so exact addresses are not exposed","make trust and safety reporting clear but not intimidating"]} | |
| {"prompt":"make me a dashboard for tracking model training runs","follow_up_prompts":["add experiments, metrics, configs, logs, artifacts, and comparison views","add alerts for failed or suspicious runs","make it easy to compare two runs side by side"]} | |
| {"prompt":"Create a compact README for a small open-source utility with installation, usage, examples, configuration, testing, and contribution notes.","follow_up_prompts":["Make the README more beginner-friendly.","Add a troubleshooting section for common setup issues."]} | |
| {"prompt":"make me a home inventory app","follow_up_prompts":["add rooms, items, photos, warranties, receipts, and search","add export for insurance purposes","make adding items faster and less tedious"]} | |
| {"prompt":"make me a web app for planning trips","follow_up_prompts":["add itinerary days, places, budgets, notes, and a packing checklist","make it detect schedule conflicts and long travel gaps","make it shareable with friends so they can suggest changes"]} | |
| {"github_repo":"eslint/eslint","prompt":"Pick one small edge case in a lint rule or config behavior, reproduce it with a test, and implement a minimal fix consistent with ESLint's existing patterns.","follow_up_prompts":["Review the change for backwards compatibility.","Add a concise note explaining the edge case."]} | |
| {"prompt":"Build a finance tracker dashboard with accounts, monthly budgets, recurring expenses, spending categories, and trend charts.","follow_up_prompts":["Add better edge cases for missing or incomplete financial data.","Make the dashboard easier to scan on mobile."]} | |
| {"prompt":"make me a small farming game","follow_up_prompts":["add crops, watering, inventory, selling, upgrades, and day cycles","make the progression feel cozy and not grindy","add save data and a better UI for managing fields"]} | |
| {"prompt":"make me a contract review workflow app","follow_up_prompts":["add uploads, clause tagging, comments, approvals, risk flags, and version history","add reviewer roles for legal, finance, and requester","make the redline/change history easy to understand"]} | |
| {"prompt":"Build a simple 2D platformer prototype in the browser with keyboard controls, hazards, collectibles, and a win condition.","follow_up_prompts":["Make the controls feel smoother and more responsive.","Add a level restart flow and better visual feedback."]} | |
| {"prompt":"Build a backend service for managing event registrations with attendees, ticket types, capacity limits, and waitlist behavior.","follow_up_prompts":["Add tests for sold-out and waitlist edge cases.","Refactor the service layer to make business rules clearer."]} | |
| {"github_repo":"expressjs/express","prompt":"Explore the Express codebase, identify one small documentation or error-message improvement that would help developers understand middleware or routing behavior, and implement it with tests if appropriate.","follow_up_prompts":["Make the explanation more precise without making it longer.","Add a focused regression test if the behavior is testable."]} | |
| {"prompt":"Build a clean personal portfolio website for a freelance photographer with a homepage, gallery, about page, contact section, and responsive mobile layout.","follow_up_prompts":["Make the visual style more premium and editorial.","Add subtle motion and improve the mobile gallery experience."]} | |
| {"prompt":"make me an employee onboarding checklist app","follow_up_prompts":["add tasks for HR, IT, manager, and new hire","add due dates, dependencies, missing access warnings, and completion history","make the first-day experience feel welcoming"]} | |
| {"github_repo":"httpie/cli","prompt":"Explore HTTPie's CLI codebase, identify one small usability improvement in error output, help text, request formatting, or configuration behavior, implement it, and update tests.","follow_up_prompts":["Make the CLI output clearer without adding noise.","Add a regression test for an edge case involving unusual input."]} | |
| {"prompt":"make me an offline-first field survey app","follow_up_prompts":["add multilingual forms, required fields, GPS points, photos, and local draft storage","add sync conflict handling when two people edit the same record","make it usable in bright sunlight on a phone"]} | |
| {"github_repo":"httpie/cli","prompt":"Trace how HTTPie parses and sends a simple request. Improve one small failure path involving invalid URLs, headers, body input, sessions, or response display.","follow_up_prompts":["Add a focused test around the failure path.","Improve the developer-facing explanation in nearby docs or comments."]} | |
| {"prompt":"make me a small crm","follow_up_prompts":["add contacts, companies, deals, tasks, and notes","add a pipeline view and reminders for stale deals","make the UX feel lightweight for a solo founder"]} | |
| {"prompt":"Create an architecture overview for a medium-sized web app explaining folders, data flow, state management, routing, and testing strategy.","follow_up_prompts":["Add a new-contributor onboarding checklist.","Make the architecture doc more concise and practical."]} | |
| {"prompt":"Build a dataset cleaning dashboard with duplicate detection, field normalization, validation errors, and approve/reject workflow.","follow_up_prompts":["Improve the triage workflow for large datasets.","Add summary metrics and better error grouping."]} | |
| {"prompt":"Build a CSV analytics app where users upload a dataset and get summary stats, missing values, charts, and simple filters.","follow_up_prompts":["Improve the UX for messy or invalid CSV files.","Add a clearer data profiling summary."]} | |
| {"prompt":"Build a webhook receiver service that validates signatures, stores events, retries failed processing, and exposes a status endpoint.","follow_up_prompts":["Harden the security and replay protection.","Add structured logging and test coverage for failure paths."]} | |
| {"prompt":"Write API documentation for a small service including endpoints, request/response examples, error codes, and authentication assumptions.","follow_up_prompts":["Add edge-case examples and validation failures.","Make the docs easier to scan quickly."]} | |
| {"prompt":"make me an app for a small clinic to manage patient intake","follow_up_prompts":["add different views for front desk staff, nurses, and providers","add consent forms, missing info warnings, and a clean handoff flow","make it handle sensitive patient data more carefully"]} | |
| {"prompt":"make me a messy spreadsheet import tool","follow_up_prompts":["add column mapping, validation, duplicate detection, preview, and rollback","add handling for inconsistent schemas and weird date formats","make errors fixable in the UI instead of just rejected"]} | |
| {"github_repo":"refinedev/refine","prompt":"Inspect one example or package in refine and improve a small edge case around loading states, error states, form validation, table filtering, or navigation.","follow_up_prompts":["Add or update tests/examples for the changed behavior.","Keep the change narrowly scoped and consistent with the package style."]} | |
| {"prompt":"build me a personal finance app","follow_up_prompts":["add budgets, recurring expenses, and a way to see where my money is going","make it handle weird edge cases like refunds and transfers","make the dashboard actually useful instead of just pretty charts"]} | |
| {"prompt":"make me a research annotation workflow","follow_up_prompts":["add documents, labels, reviewers, disagreements, adjudication, and export","add quality checks for inconsistent annotations","make keyboard shortcuts and review queues efficient"]} | |
| {"prompt":"make me a data retention dashboard","follow_up_prompts":["add datasets, retention rules, legal holds, deletion approvals, and export logs","add warnings before destructive actions","make the policy lifecycle understandable to non-lawyers"]} | |
| {"prompt":"make me a 3d minecraft game","follow_up_prompts":["make it so i can play with my friends online","add crafting, inventory, and a couple different block types","make the world bigger without making it lag like crazy"]} | |
| {"prompt":"make me a wordle clone"} | |
| {"github_repo":"tastejs/todomvc","prompt":"Explore the repository structure, pick one actively maintained example app, identify one small usability or consistency improvement, implement it, and update any relevant tests or documentation.","follow_up_prompts":["Refactor the changed area to better match the style of the surrounding example.","Add a short note explaining the behavior you changed and how to verify it."]} | |
| {"prompt":"make me a simple deployment dashboard","follow_up_prompts":["add services, environments, deploy history, health checks, logs, and rollback buttons","add scary confirmation flows for risky actions","make failure states obvious and useful"]} | |
| {"prompt":"make me a feature flag dashboard","follow_up_prompts":["add flags, environments, rollout percentages, targeting rules, and audit history","add safeguards so i don't accidentally enable something for everyone","make the rule builder easier to understand"]} | |
| {"github_repo":"fastapi/full-stack-fastapi-template","prompt":"Inspect the frontend portion of this template and improve one small UX state, such as loading, empty, error, auth, or form validation behavior.","follow_up_prompts":["Make the improved state responsive and accessible.","Add or update tests where the project structure supports it."]} | |
| {"prompt":"Build a CSV cleanup tool that normalizes column names, trims whitespace, detects malformed rows, and writes a cleaned output file.","follow_up_prompts":["Add a validation report and configurable rules.","Improve error messages and test coverage."]} | |
| {"prompt":"make me a procurement dashboard for a city department","follow_up_prompts":["add planning, bids, vendors, awards, contracts, spending, and implementation status","add red flag indicators for unusual procurement patterns","make the public transparency view different from the internal admin view"]} | |
| {"prompt":"make me an org chart and access review app","follow_up_prompts":["add employees, managers, teams, systems, permissions, and review decisions","add graph-style relationship views and audit history","make risky access stand out without exposing too much HR data"]} | |
| {"prompt":"make me a literature review manager","follow_up_prompts":["add papers, tags, notes, quotes, citations, related work, and reading status","add duplicate detection and PDF metadata cleanup","make the relationship map between papers useful"]} | |
| {"prompt":"make me a physical therapy exercise tracker","follow_up_prompts":["add exercise plans, reps, pain ratings, progress notes, and therapist comments","add video/photo placeholders and patient adherence summaries","make it easy for older users with limited mobility"]} | |
| {"prompt":"make me a property inspection app","follow_up_prompts":["add rooms, checklist items, photos, damage notes, signatures, and repair follow-ups","add offline mode for inspections with bad reception","make the final report look professional"]} | |
| {"prompt":"make me a calendar scheduling app","follow_up_prompts":["add availability blocks, meeting types, booking links, and timezone handling","add rescheduling and cancellation flows","make the timezone stuff very obvious so users don't get confused"]} | |
| {"prompt":"make me a policy attestation app","follow_up_prompts":["add policies, employee acknowledgements, reminders, exceptions, and reports","add audit evidence for compliance reviews","make it handle people who join or leave mid-cycle"]} | |
| {"prompt":"make me a fleet maintenance tracker","follow_up_prompts":["add vehicles, mileage, inspections, repairs, service intervals, costs, and downtime","add alerts for overdue maintenance and recurring issues","make the dashboard useful for a fleet manager"]} | |
| {"prompt":"make me a route planning app for volunteers delivering supplies","follow_up_prompts":["add addresses, availability windows, driver capacity, priority requests, and route maps","add last-minute cancellation and reassignment flows","make it simple enough for non-technical volunteers"]} | |
| {"prompt":"Build me a Python library for parsing and querying large JSONL datasets with a fluent API: lazy streaming, filtering, projection, group-by, and writing back out. Make it memory-efficient on multi-GB files and mistake-free.","system":"You are generating a fresh, self-contained project. Choose a sensible minimal implementation, include clear run instructions and tests, avoid placeholder-only code, and prioritize correctness over scope.","follow_up_prompts":["Add a CLI wrapper around the library with sensible defaults.","Add one performance optimization and show its effect on a large file."]} | |
| {"prompt":"Build me an async Python web scraper with rate limiting, retries with backoff, concurrency limits, and a resumable on-disk queue. Pick a sensible stack and make no mistakes.","system":"Favor a production-shaped design: validate inputs, handle failures consistently, separate fetching from parsing, and include focused tests.","follow_up_prompts":["Add polite robots.txt handling and per-domain rate limits.","Add structured logging and a progress summary."]} | |
| {"prompt":"Build me a Rust CLI that watches a directory and incrementally builds a search index over text files, with fast substring and fuzzy queries from the terminal. Keep it focused, fast, and mistake-free.","system":"You are generating a fresh, self-contained project. Choose a sensible minimal implementation, include clear run instructions, avoid placeholder-only code, and prioritize correctness over scope.","follow_up_prompts":["Add incremental re-indexing on file change instead of full rebuilds.","Add one ranking improvement to the query results."]} | |
| {"prompt":"Build me a small Rust HTTP service using axum with typed request/response models, structured error handling, graceful shutdown, and a couple of integration tests. Pick a realistic domain yourself.","system":"Prioritize idiomatic Rust: clear error types, no unwrap in request paths, and separation between handlers and business logic.","follow_up_prompts":["Add request validation and consistent JSON error responses.","Add one endpoint that a real user of this service would need."]} | |
| {"prompt":"Build me a TypeScript library for typed, retryable fetch with timeouts, request deduplication, and a small response cache, exposed as a clean ergonomic API. Make it tree-shakeable and mistake-free.","system":"Favor strong types, zero dependencies where reasonable, and focused tests for edge cases (timeouts, aborts, cache invalidation).","follow_up_prompts":["Add an exponential-backoff retry policy with jitter.","Add one developer-experience improvement to the API."]} | |
| {"prompt":"Build me a Node + TypeScript CLI that generates a typed API client from an OpenAPI spec. Pick sensible defaults, include dry-run output, and make no mistakes.","system":"Design this like a real developer tool. Include safe defaults, clear error output, dry-run behavior, and tests for tricky spec shapes (refs, enums, nullable fields).","follow_up_prompts":["Add support for one more spec edge case (oneOf or discriminated unions).","Add a flag a developer would actually use."]} | |
| {"prompt":"Build me a SvelteKit app with a small backend for a personal bookmarking tool: add, tag, search, and archive links, with server-side load functions and form actions. Pick the details yourself and make no mistakes.","system":"You are generating a fresh, self-contained project. Use idiomatic SvelteKit (load functions, form actions, progressive enhancement), keep it runnable, and avoid placeholder-only code.","follow_up_prompts":["Add optimistic UI for adding and archiving.","Add one server-side feature that makes it feel complete."]} | |
| {"prompt":"Build me a Svelte component library: a few polished, accessible primitives (button, dialog, combobox, toast) with keyboard support and a small docs page demonstrating them.","system":"Prioritize accessibility (focus management, ARIA, keyboard) and a clean, composable component API.","follow_up_prompts":["Add a theming system with CSS variables.","Add one more primitive that pairs well with the existing set."]} | |
| {"prompt":"Build me a React dashboard using shadcn/ui and Tailwind for a small analytics app: cards, a data table with sorting and filtering, a date-range picker, and a couple of charts. Make it look clean and professional and make no mistakes.","system":"Use shadcn/ui components idiomatically, keep the layout responsive and accessible, and prioritize a coherent, modern look.","follow_up_prompts":["Add loading, empty, and error states throughout.","Add one interaction that makes the dashboard feel real."]} | |
| {"prompt":"Build me a React multi-step form with shadcn/ui, react-hook-form, and zod validation: per-step validation, a review step, and clear inline errors. Make it polished and mistake-free.","system":"Favor accessible, well-typed forms: schema-driven validation, focus management on errors, and no uncontrolled-input mistakes.","follow_up_prompts":["Add autosave of in-progress form state.","Add one quality-of-life feature for long forms."]} | |
| {"prompt":"Build me a Next.js (App Router) app with server components, route handlers, and server actions for a small link shortener: create short links, redirect, and show click stats. Make reasonable decisions yourself and make no mistakes.","system":"Use idiomatic App Router patterns (server components by default, server actions for mutations), keep it runnable, and handle errors and not-found states cleanly.","follow_up_prompts":["Add edge-friendly caching for redirects.","Add one feature a real user of a link shortener would want."]} | |
| {"prompt":"Build me a Cloudflare Workers API in TypeScript backed by D1 and KV for a small URL-metadata service: store, fetch, and cache page metadata, with sensible cache TTLs and consistent error responses. Pick the details yourself and make no mistakes.","system":"Favor a production-shaped Workers design: validate inputs, return consistent errors, use KV and D1 idiomatically, and include focused tests with the Workers test tooling.","follow_up_prompts":["Add per-client rate limiting.","Add a scheduled cron worker that refreshes stale metadata."]} | |
| {"prompt":"Take a Next.js app and configure it to deploy on Cloudflare with OpenNext: set up the OpenNext Cloudflare adapter, wrangler config, and bindings, and make the build plus a server-action route and a cached route work end to end. Explain the key config choices.","system":"You are doing real deployment engineering. Keep config minimal and correct, explain bindings and runtime constraints, and call out anything that differs from a Node deployment.","follow_up_prompts":["Add an R2 or KV binding used by one route.","Document a likely deployment failure and how to fix it."]} | |
| {"prompt":"Build me a Jupyter notebook that trains and evaluates a small text classifier end to end: load data, EDA with charts, a baseline, a stronger model, error analysis, and a clear conclusion. Make it run top-to-bottom without errors on a fresh kernel.","system":"You are producing a clean, reproducible notebook. Keep cells ordered so a fresh kernel runs top-to-bottom, and explain decisions in markdown between cells.","follow_up_prompts":["Add proper cross-validation instead of a single split.","Add one visualization that makes the error analysis clearer."]} | |
| {"prompt":"Build me a marimo notebook (reactive Python) that's an interactive data explorer: file upload, UI controls for filtering and grouping, and charts that update reactively as inputs change. Make it run cleanly and use marimo's reactivity correctly.","system":"You are building a reactive marimo notebook. Use marimo UI elements and the reactive dataflow correctly (no hidden global mutation), and keep the app responsive.","follow_up_prompts":["Add a derived-metric panel that recomputes reactively.","Add an export of the current filtered view."]} | |
| {"prompt":"Set me up a pnpm workspaces monorepo with two packages (a shared TS library and a React app that consumes it), workspace-protocol linking, a shared tsconfig with project references, and build/test scripts that run across the workspace. Make it work end to end.","system":"You are doing real project-setup engineering. Keep the workspace config correct, ensure the app actually consumes the library, and make scripts runnable.","follow_up_prompts":["Add changesets-style versioning for the packages.","Show the equivalent setup differences for npm and yarn workspaces."]} | |
| {"prompt":"Write me a script and runbook to migrate a JavaScript project from npm to pnpm safely: lockfile conversion, npm-specific config handling, peer-dependency differences, and a verification step. Include a dry-run and rollback guidance.","system":"Design this like a real developer tool and runbook. Include safe defaults, a dry-run, clear output, explicit verification, and call out the gotchas (hoisting, lockfile, scripts).","follow_up_prompts":["Add handling for one tricky case (patched dependencies or postinstall scripts).","Add a check that the install graph matches before and after."]} | |
| {"prompt":"Build me a Solidity ERC-20 staking contract as a Foundry project: stake, withdraw, time-based rewards, and an emergency pause, plus a thorough test suite covering the reward math and edge cases. Make it correct and mistake-free.","system":"You are doing smart-contract engineering. Prioritize correctness and safety: checks-effects-interactions, no reentrancy, careful reward accounting, and high-coverage Foundry tests.","follow_up_prompts":["Add a fuzz test for the reward accounting.","Add one feature a real staking contract would need, like updatable reward rates."]} | |
| {"prompt":"Build me a Solidity + Foundry project for a minimal on-chain name-to-address registry with owner-controlled updates and resolution, plus tests and gas snapshots. Pick reasonable design choices and make no mistakes.","system":"You are doing smart-contract engineering. Favor a clean, gas-aware design, explicit access control, and a complete Foundry test suite including revert cases.","follow_up_prompts":["Add events and an indexing-friendly interface.","Add a gas-optimization pass and report the savings."]} | |
| {"prompt":"Build me a Vyper escrow contract with deposit, release, and refund, time-locked, with a clear state machine, plus tests. Keep it idiomatic Vyper and mistake-free.","system":"You are doing smart-contract engineering in Vyper. Favor explicit state, safe arithmetic, careful access control, and tests covering each state transition and revert path.","follow_up_prompts":["Add a dispute/timeout path and test it.","Compare one design decision against how you'd do it in Solidity."]} | |
| {"prompt":"Build me a small TypeScript dApp frontend with viem and React that connects a wallet, reads an ERC-20 balance, and submits a transfer with proper transaction state handling (pending, confirmed, error). Make the UX clean and mistake-free.","system":"You are doing real dApp engineering. Handle wallet connection, chain switching, and the transaction lifecycle correctly, and keep reads and writes typed.","follow_up_prompts":["Add gas estimation and a confirmation step before sending.","Add support for one more token interaction."]} | |
| {"prompt":"Implement Group Relative Policy Optimization (GRPO) from scratch for a small language model. Sample a group of completions per prompt, compute group-relative advantages without a value network, apply the clipped surrogate objective with a KL penalty to a frozen reference, and log reward, KL, and advantage statistics per step. Get it training stably on a single GPU.","system":"You are doing ML research engineering. Prioritize a correct, readable, runnable implementation over scale. Use a tiny model and a toy verifiable-reward task so the loop runs end to end.","follow_up_prompts":["Add a deterministic RLVR-style reward (exact-match on a math task) and show clean reward and KL curves.","Add one practical training-stability improvement that fits your implementation."]} | |
| {"prompt":"Build a knowledge distillation pipeline that distills from multiple teacher models into a smaller student (e.g. Qwen3 or Gemma). Support token-level KL distillation on logits and sequence-level distillation on sampled traces, handle tokenizer mismatch between teachers and student, and let me weight or route between teachers per example.","system":"You are doing ML research engineering. Favor correctness and clear configuration. Handle vocab/tokenizer alignment explicitly, keep memory bounded for large logit tensors, and make the teacher-aggregation strategy swappable.","follow_up_prompts":["Add temperature and top-k logit truncation for the soft targets and pick sensible defaults.","Add a held-out eval that reports student-vs-teacher agreement against a no-distillation baseline."]} | |
| {"prompt":"Implement on-policy distillation: have the student generate completions, score them with a teacher, and update the student toward the teacher's distribution on the student's own samples. Compare it against offline sequence-level distillation and show where each wins.","system":"You are doing ML research engineering. Keep the training loop inspectable, use a small model and toy task, and report the metrics that distinguish on-policy from offline distillation.","follow_up_prompts":["Add a KL-to-teacher and a self-entropy metric and explain what each tells you."]} | |
| {"prompt":"Build a reproducible evaluation harness for a local Qwen model. Load tasks from JSONL, run batched generation with fixed seeds, apply pluggable deterministic scorers (exact-match, regex, sandboxed unit-test execution, and an LLM-judge), and emit per-sample traces plus aggregate metrics with confidence intervals.","system":"Build this as a real research tool. Separate task loading, generation, and scoring; make runs reproducible (seeds, decoding params, prompt hashing); and treat partial failures as data, not crashes.","follow_up_prompts":["Add pass@k estimation with the unbiased estimator and bootstrap confidence intervals.","Add a clean HTML or terminal report that makes per-task results easy to scan."]} | |
| {"prompt":"Build a reliable LLM-as-judge scorer for open-ended model outputs: support pointwise rubric scoring and pairwise comparison, randomize answer order, parse judge output robustly, and calibrate against a small human-labeled set. Report judge-human agreement.","system":"You are doing eval research. Make the judge reliable and its output easy to parse and aggregate; keep prompts and rubrics explicit.","follow_up_prompts":["Add a tie-handling policy and an ensemble-of-judges mode.","Add caching so repeated comparisons don't re-call the judge."]} | |
| {"prompt":"Build a trajectory grader for agent traces: given a JSONL log of tool calls, observations, and final outputs, score each trajectory on task success, efficiency (tokens and steps), and tool-call validity, then build a best-of-N reranker that selects the strongest trajectory for each task.","system":"You are doing harness/eval research. Parse traces by structure, not regex; make the scoring rubric explicit and inspectable; and include a no-reranker baseline.","follow_up_prompts":["Add early-stopping that ends a trajectory once it's clearly going to succeed, and measure compute saved."]} | |
| {"prompt":"Implement speculative decoding with a small draft model and a larger target model: draft k tokens, verify them in a single target forward pass, accept the longest valid prefix, and resample on rejection. Measure tokens-accepted-per-step and wall-clock speedup vs the target alone.","system":"You are doing inference research. Use a correct acceptance rule that preserves the target distribution, and make the acceptance statistics observable.","follow_up_prompts":["Add a tunable draft length and plot the speedup vs acceptance-rate tradeoff."]} | |
| {"prompt":"Implement a KV cache for autoregressive decoding and extend it to a simple paged-attention-style block allocator supporting multiple concurrent sequences with continuous batching. Show memory usage and throughput vs a naive per-sequence cache.","system":"You are doing inference systems research. Keep the allocator logic clear and correct under variable-length sequences; measure the throughput win.","follow_up_prompts":["Add prefix sharing for a common system prompt across sequences and quantify the savings."]} | |
| {"prompt":"Implement post-training weight quantization for a small transformer: int8 and int4 with per-channel scales, a calibration pass on a small dataset, and an evaluation of the perplexity/accuracy vs memory tradeoff against an fp16 baseline.","system":"You are doing efficiency research. Keep calibration and evaluation reproducible and report the per-layer error.","follow_up_prompts":["Add a mixed-precision policy that keeps sensitive layers in higher precision and justify the choice."]} | |
| {"prompt":"Implement and compare multi-head, multi-query, and grouped-query attention in one small transformer with a clean switch between them. Measure quality, KV-cache size, and decode latency for each, and explain the tradeoffs.","system":"You are doing architecture research. Keep the implementations side-by-side and numerically checkable; report cache size and latency alongside loss.","follow_up_prompts":["Add a rough multi-head latent attention (MLA) variant and discuss where it helps."]} | |
| {"prompt":"Train a byte-pair-encoding tokenizer on a small corpus and analyze it: compression ratio, fertility (tokens per word) across languages and code, handling of digits and whitespace, and the byte-fallback rate. Compare against an existing pretrained tokenizer.","system":"You are doing ML research. Make the analysis quantitative and reproducible, and include the interesting cases (code indentation, non-Latin scripts).","follow_up_prompts":["Add a digit-tokenization experiment and discuss its effect on arithmetic."]} | |
| {"prompt":"Build a near-duplicate detection and deduplication pipeline for a text training set using MinHash + LSH, with configurable similarity thresholds, shard-parallel processing, and a report of how much data each stage removes. Include exact-dup hashing as a fast first pass.","system":"You are building a real data tool. Keep it streaming/shardable for large corpora, make removal decisions auditable, and include focused tests on known duplicate and near-duplicate pairs.","follow_up_prompts":["Add cross-split decontamination that removes training docs overlapping a held-out eval set.","Add a sample-review mode that surfaces borderline pairs near the threshold."]} | |
| {"prompt":"Train a sparse autoencoder on the residual-stream activations of a small language model to extract interpretable features. Report reconstruction loss, L0 sparsity, and dead-feature rate, and build a small viewer that shows top-activating examples per feature.","system":"You are doing interpretability research. Cache activations efficiently, keep the SAE training loop inspectable, and make the resulting features browsable.","follow_up_prompts":["Add an activation-patching experiment that tests whether a discovered feature is causal for a behavior."]} | |
| {"prompt":"Implement a logit lens and a set of linear probes for a small transformer: decode intermediate residual streams into vocabulary space layer by layer, and train probes to detect when a factual answer becomes linearly decodable. Visualize where in the network the answer forms.","system":"You are doing interpretability research. Keep hooks and caching clean, and make the per-layer results easy to read.","follow_up_prompts":["Add a probe-generalization test across paraphrases of the same fact."]} | |
| {"prompt":"Build a synthetic instruction-data generation pipeline: generate candidate (instruction, response) pairs from a teacher model, then filter with rule checks, an LLM critic, deduplication, and difficulty/diversity scoring. Output a clean SFT-ready JSONL with provenance for every example.","system":"You are doing data research. Make the filtering stages modular and measurable, and keep provenance for auditability.","follow_up_prompts":["Add a self-consistency filter and measure its effect on downstream eval.","Add a diversity sampler so the final set isn't dominated by a few templates."]} | |
| {"prompt":"Train a pairwise preference reward model on a small dataset: build the data loader, the Bradley-Terry-style loss, and an evaluation that reports preference accuracy and score calibration on a held-out split. Make the training run cleanly end to end.","system":"You are doing alignment research engineering. Keep the model and training loop small and inspectable, and include a length-controlled evaluation so scores aren't just tracking length.","follow_up_prompts":["Add a length-debiasing term and re-measure accuracy.","Add a simple API that scores a new (prompt, response) pair."]} | |
| {"prompt":"Implement RoPE positional encoding and experiment with context-length extension methods (position interpolation and NTK-aware scaling) on a small model. Measure perplexity as context grows past the training length and show where each method degrades.","system":"You are doing architecture research. Keep the RoPE math correct and the extrapolation experiment reproducible; report perplexity vs position.","follow_up_prompts":["Add a needle-in-a-haystack retrieval probe across context positions."]} | |
| {"prompt":"Implement a small mixture-of-experts feed-forward layer with top-k token routing, a load-balancing auxiliary loss, and capacity factors with token dropping. Instrument expert utilization and show how the balance loss changes routing.","system":"You are doing architecture research. Make routing decisions and expert utilization observable, and keep the implementation correct under token dropping.","follow_up_prompts":["Add expert-choice routing as an alternative and compare utilization."]} | |
| {"prompt":"Build a trajectory-collection harness that runs an agent against a set of tasks, captures full traces (prompts, tool calls, observations, rewards) to JSONL, supports resuming after crashes, and deduplicates identical rollouts. Make it a solid data source for SFT or RL post-training.","system":"You are building a real research data pipeline. Make traces self-describing and replayable, handle partial/crashed runs gracefully, and keep the schema stable enough to train on.","follow_up_prompts":["Add a filtering pass that keeps only high-reward trajectories with configurable criteria.","Add a replay tool that re-scores old traces with a new reward function."]} | |
| {"prompt":"Build a decoding-strategy testbed for a local model: greedy, temperature, top-k, top-p, min-p, and beam search, with a fixed eval harness measuring quality, diversity (distinct-n, self-BLEU), and latency for each. Produce a tradeoff table.","system":"You are doing inference research. Keep sampling implementations correct and seeded, and measure diversity and quality together.","follow_up_prompts":["Add a contrastive or speculative variant and place it on the same tradeoff table."]} | |
| {"prompt":"Build a training-run logger and dashboard for fine-tuning runs that tracks loss, grad norm, learning rate, token throughput, GPU memory, and eval metrics, detects divergence/NaNs early, and supports comparing multiple runs. Make it backend-agnostic (local files or a tracking server).","system":"You are building a real research tool. Prioritize early detection of failed runs, reproducible run metadata, and clean multi-run comparison.","follow_up_prompts":["Add a clean run-comparison view that overlays metrics from several runs."]} | |
| {"github_repo":"walkinglabs/hands-on-modern-rl","prompt":"Pick one runnable RL example in the code/ directory (PPO, DPO, or GRPO), reproduce its training run, find one correctness or reproducibility issue (seeding, metric computation, advantage normalization, or a broken default), fix it, and verify the training curve still behaves sensibly.","system":"You are working inside an existing open-source repository. Keep the patch narrowly scoped, follow the project style, add or update focused checks when practical, and avoid broad rewrites.","follow_up_prompts":["Add a short note to the chapter docs explaining the fix and the expected metric behavior.","Make the example deterministic given a fixed seed and document any remaining nondeterminism."]} | |
| {"github_repo":"ColeMurray/background-agents","prompt":"Trace how the automations system handles repeated failures, and improve one operational edge case: the auto-pause-after-failures policy, dead-letter behavior for permanently failing webhook triggers, or context handling when a long-running session is resumed from a snapshot. Keep public behavior stable.","system":"You are working inside an existing open-source repository. Keep the patch narrowly scoped, follow the project style, add or update focused tests when practical, and avoid broad rewrites.","follow_up_prompts":["Add a focused test for the failure path you changed.","Document the operational behavior for someone deploying single-tenant."]} | |
| {"github_repo":"AletheiaResearch/mnemosyne","prompt":"Inspect how mnemosyne anonymizes conversation histories during extract/attest, and improve one edge case in the identity/entity scanning or redaction path (a PII pattern it misses, or an over-redaction that destroys useful structure). Keep output local-only behavior intact.","system":"You are working inside an existing open-source repository. Keep the patch narrowly scoped, follow the project style, add or update focused tests when practical, and avoid broad rewrites.","follow_up_prompts":["Add a test fixture with a tricky redaction case and verify the scan handles it.","Document the redaction guarantee and its known limits."]} | |
Xet Storage Details
- Size:
- 89.3 kB
- Xet hash:
- 374f786d50b77d745a9d7ceaf7f6c632ab0e10ab462bf03d0eb17d2c9a208145
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.