AI prompt builder

Nextron already generates AGENTS.md and CLAUDE.md with project-specific conventions. This builder adds the missing task brief: what the agent should accomplish, which generated stack it must preserve, what it should inspect first, and how it must verify the result.

Prompt builder

Give your coding agent the whole brief.

Choose a task, load the generated stack, and copy a prompt grounded in Nextron’s actual architecture.

Choose the task

Start from a tested Nextron workflow.

Prompt recipe

Map the generated stack, module boundaries, routes, and development workflow.

Load your stack

Your configuration stays in this browser.

nextron.config.ts

Review or enter the stack manually

Project nameProject typePackage managerORMDatabaseDatabase providerAuthenticationAI provider

Describe the outcome

Specific inputs produce a more useful implementation.

RequirementsConstraints OptionalExpected result Optional

Generated prompt

Ready for your coding agent

You are working inside a Nextron-generated application. Treat the repository as the source of truth and do not rely on remembered framework APIs when local code or current documentation is available.

## Objective

Audit this Nextron project and explain how its generated architecture works before proposing changes.

## Requested work

Inspect the repository and infer the concrete requirements from the requested task before editing.

## Detected stack

- Project: my-app (app)
- Framework: Next.js App Router with strict TypeScript and Tailwind CSS v4
- Package manager: pnpm
- Database: postgresql via local; ORM: drizzle
- Authentication: better-auth
- Optional modules: dashboard

## Inspect before changing code

1. Read AGENTS.md or CLAUDE.md completely.
2. Read nextron.config.ts and package.json.
3. Inspect the relevant route under src/app and its implementation under src/modules.
4. Check src/env.ts before introducing or reading environment variables.
5. Reuse existing dependencies and conventions unless the task explicitly requires a change.

## Nextron architecture constraints

- Keep App Router files thin; page-level UI belongs in src/modules/[feature]/ui/views.
- Put mutations in src/modules/[feature]/server/actions.ts and read-only queries in server/queries.ts.
- Put module-specific types in src/modules/[feature]/@types/index.ts.
- Default to Server Components and add 'use client' only for browser state or interaction.
- Access validated environment values through src/env.ts, never process.env directly.
- Use provider-specific APIs for the configured authentication, database, and AI stack.

## Additional constraints

Preserve existing behavior outside the requested scope and do not replace generated infrastructure without evidence.

## Expected result

- Identify the installed stack from repository evidence.
- Map routes, modules, server boundaries, and environment configuration.
- Call out inconsistencies between configuration, documentation, and source code.

## Verification

- Run the smallest relevant checks while iterating.
- Before finishing, run pnpm typecheck and pnpm build when those scripts exist.
- Report changed files, verification results, and any remaining setup that requires secrets or external services.

## Current Nextron references

- Documentation index: https://nextron.dev/llms.txt
- Complete documentation context: https://nextron.dev/llms-full.txt
- Relevant guide: https://nextron.dev/docs/project-structure

Do not stop at a proposal. Implement and verify the requested work unless a destructive action, missing secret, or product decision genuinely requires confirmation.

2,642 characters · edit your inputs to regenerate

How the builder uses your configuration

Paste the generated nextron.config.ts file to fill the stack selectors automatically. Parsing happens entirely in your browser. The builder recognizes Nextron’s JSON-shaped defineConfig(...) output and never evaluates the pasted code.

If the file has been customized into arbitrary TypeScript, use the manual selectors. Always review the detected stack before copying the prompt.

Give the agent current context

Every generated prompt directs the agent to inspect repository files before editing and includes links to:

/llms.txt for the documentation index.
/llms-full.txt for the complete documentation context.
The Nextron guide most relevant to the selected task.

The prompt is deliberately tool-neutral. Use it with any coding agent that can inspect and modify your repository.