Skip to main content
Never paste secrets such as SUPABASE_SERVICE_ROLE_KEY, OPENAI_API_KEY, ANTHROPIC_API_KEY, MIDTRANS_SERVER_KEY, DOKU_SECRET_KEY, or other production credentials into an AI tool.
Replace every placeholder in square brackets such as [describe your product] before sending the prompt. If the AI tool has access to your repository, let the prompt tell it to inspect the relevant files first. If it does not have repo access, ask it to state assumptions clearly.

How to use the prompts on this page

  1. Pick the prompt closest to your task.
  2. Add real context about the product, active features, and launch goal.
  3. Ask for structured output: steps, relevant files, needed env vars, risks, and verification.
  4. Cross-check the answer against this documentation before making large changes.

Base instruction block

If you want more consistent AI output, prepend this block to almost any prompt:

Prompts for non-technical users

Use this when: you are not yet sure which KilatKoding scenario matches your product.Expected output: best-fit use case, active modules, minimum services, and mismatch risks.
Use this when: you are still unsure whether this boilerplate is the right starting point.Expected output: recommended path, reasons, trade-offs, and risk signals.
Use this when: you know what you want to build but do not know how to explain it clearly to a developer or coding assistant.Expected output: a structured brief with active features, integrations, and custom work.
Use this when: you want AI to translate the launch docs into a checklist that fits your actual setup.Expected output: a launch checklist that both non-technical and technical teammates can use.
Use this when: you want a safe boundary between non-technical work and technical implementation.Expected output: a clear split between non-code tasks and developer tasks.

Prompts for developers

Use this when: you want a more prescriptive .env.local and service setup starting point.Expected output: env vars, toggles, services, and setup order.
Use this when: you want AI to summarize local setup for a cross-platform team.Expected output: setup steps, dependencies, and verification checkpoints.
Use this when: you want to turn off auth, payments, admin, contact, waitlist, or AI.Expected output: toggle changes, files to inspect, and UI/copy cleanup tasks.
Use this when: you want to add a product-specific feature without breaking the existing foundation.Expected output: phased implementation plan, affected files, and integration risks.
Use this when: you need directed diagnosis instead of generic debugging advice.Expected output: likely root causes, check order, and safe fixes.

Prompts for agencies or freelancers

Use this when: you receive a client brief and want a quick fit assessment.Expected output: fit analysis, custom work areas, and scope-creep signals.
Use this when: you want to turn KilatKoding into a client project without leaving confusing leftovers.Expected output: rebranding plan, active modules, and cleanup list.
Use this when: you want to explain clearly what is already included and what needs additional implementation.Expected output: a proposal-friendly scope split.
Use this when: the project is almost done and you want a cleaner client handoff.Expected output: handoff notes, ownership boundaries, and post-handoff tasks.
Use this when: you want a final sanity check before client delivery or production launch.Expected output: final checks, blind spots, and residual risks.

Docs pages that pair well with these prompts

Use cases

Use the scenario-selection prompt if you are still deciding which path fits the product best.

Comparison

Use the starting-point evaluation prompt if you are still deciding between a boilerplate, UI template, or custom build.

Preset recipes

Use the env and toggle prompt if you want a more prescriptive .env.local starting point.

Customization

Use the rebranding and file-mapping prompt if you want to change KilatKoding with less guesswork.

Launch checklists

Use the launch audit prompt if you need a checklist tailored to your live product setup.

Troubleshooting

Use the diagnosis prompt when you are dealing with bugs, fallback behavior, or incomplete setup.
If you want better AI output, include concrete context: use case, active features, providers in use, your latest changes, and the exact output you want. More specific prompts almost always produce more useful answers.