Tested prompts for building with Claude
Prompts that help you build, ship and run small AI-powered web apps. Each one is tested, explained and free to reuse with credit. Copy it, fill in the blanks, run it.
/
25 prompts · 7 with a SKILL.md · add yours
- App specsPrompt Spec a maturity assessment app: content contract, tested scoring, results in the URL Spec a browser-only maturity assessment whose content is checked data and whose scoring is pinned by simulated respondents, so every result is reproducible from a link. advanced
- App specsPrompt Spec an AI app's inputs, outputs and model stages before any code Turn a one-paragraph AI app idea into a spec of every input, output, model call, code check, cost and fallback, so the model's job stays narrow and testable. intermediate
- AI featuresPrompt Ask 2–4 clarifying questions with tappable options Before an expensive generation, ask only the questions whose answers would change the result, as JSON your UI can render as chips. beginner
- AI featuresPrompt Model proposes, code disposes: split one big LLM call into checked stages Let the model return only choices and unit values while code derives every total, checks each unit and re-prompts only the parts that fail. advancedSKILL.md
- AI featuresPrompt Parallel specialist research agents, then a composer that marks unverified facts Fan research out to narrow specialists, each with its own search budget and word cap, then compose one result that flags every unsourced fact. intermediate
- AI featuresPrompt Stream generated pages as data, not markup, with an NDJSON block vocabulary Have the model write a page as one JSON block per line from a fixed vocabulary, so your app renders it as it streams, themes it freely and never renders model-written HTML. advanced
- Claude CodeClaude Code Allow, ask, deny: Claude Code permissions and a SessionStart status hook Sort a repo's commands by what they can break: routine work runs unprompted, merges wait for an Allow tap, pushes to main and deploys are refused; a start hook flags a broken toolchain. intermediate
- Claude CodePrompt Custom instructions for senior builders: one recommendation, conclusion first, no padding Make Claude state the problem first, commit to one diagnosed fix with a check per step, and leave out what a senior engineer already knows. beginner
- Claude CodeSkill Publish from your phone: checkpoints, a facts gate and a one-tap merge Claude Code researches, drafts and publishes a post from your phone, stopping for your tap at the outline, review and merge, and counting a number as verified only when a fetched quote backs it. advancedSKILL.md
- Claude CodeClaude Code Write CLAUDE.md as a short, checkable contract from the repo's scripts and CI Claude Code reads the repo's scripts, CI and docs and writes a short CLAUDE.md in which every rule cites a source and can be checked in review. beginner
- Claude CodePrompt Write PR descriptions with evidence, debatable choices and what wasn't tested Turn a diff into a PR description a reviewer can act on, with the defect in numbers, what changed per file, the choices they might dispute, what wasn't tested and how to undo it. beginner
- DesignClaude Code Art bible plus a build-blocking SVG contract, so any AI tool can add art Split your art rules into a style guide people judge by eye and an SVG contract a script checks before every build, so AI-drawn art can't collide, leak or break the sprite. intermediate
- DesignClaude Code One-screen mobile layout: no page scroll, proven by a Playwright walk Fit every screen of a flow into one viewport (top bar, scaling middle, bottom action bar), then walk it at several sizes and get a table of anything that scrolls, overflows or gets clipped. intermediate
- DesignPrompt Pick a visual direction by bake-off, then write it up as DESIGN.md Build 3–4 named specimen pages, judge them against your product's jobs, graft the best ideas from the losers, and write a DESIGN.md with a colour budget and AA fixes. intermediate
- DesignPrompt Reality-checked briefs for every figure, before anyone draws it Write one checkable brief per figure before anything is drawn, covering the real subject, the facts the picture has to get right, how faithful it claims to be and the mistakes to avoid. beginner
- DesignPrompt Turn your design system into a reference-first Claude skill Turn your tokens and brand notes into a Claude skill: a short SKILL.md that orients, one tokens file that holds every value, and component files that load on demand. intermediateSKILL.md
- Deploy & hostingClaude Code Build a launch pad: one push gives an app its own repo, Worker and subdomain Claude Code builds a monorepo where pushing one app folder makes a GitHub Action create its repo, deploy it to its own subdomain and wire auto-deploy, refusing anything it didn't create. advancedSKILL.md
- Deploy & hostingSkill Hardening round for a public AI app that spends your own API key Before launch, audit and patch a public AI app that spends your API key: fail-closed quotas, keyed IP hashes, a bot gate, CSP, provider terms, spend-cap arithmetic and bring-your-own-key. advancedSKILL.md
- Deploy & hostingClaude Code Risky one-off infra change: check mode, typed apply, automatic rollback Have Claude Code write a DNS cutover or migration as a workflow that proves every permission on throwaway probes, refuses ambiguous matches and rolls itself back. advanced
- QA & debuggingClaude Code A local mock LLM API where a typed word triggers each failure Claude Code builds a dev fake of your LLM API that streams the real wire format and fails on cue, from a 401 to a cut stream, when a trigger word is typed into the app. intermediate
- QA & debuggingClaude Code Make the privacy note and credits match the code Claude Code inventories every data flow from the code, with file and line, rewrites the privacy note and credits to match, and leaves a review marker that CI warns about. intermediate
- QA & debuggingPrompt Source-tight copy audit: every claim says no more than its source Make every number in your copy say no more than its source: quote the passage, label the figure, then keep, tighten or cut it. intermediateSKILL.md
- QA & debuggingClaude Code Turn an adversarial review into acceptance checks and one regression test per defect Claude Code turns each defect from a review of one AI output into a check that runs on every result, a mock-model replay and a named offline regression test. advanced
- Data & geospatialPrompt QGIS layer builder: the open-data source, its CRS and exact steps to load it Turn a layer request and a place into the best open source, its access terms and CRS, exact QGIS menu steps and a URL or query to paste. intermediateSKILL.md
- Data & geospatialPrompt Spec a no-backend site-planning studio POC on free, keyless map data Turn a site-planning idea into a POC spec for a static map app, where constraints with provenance drive a rules-based layout and every free data source has a fallback. advanced
No prompts match.