[Audits Aren't Optional]
[PLATFORM SAFETY GUARANTEE]
> How to Trigger an Audit
Use this prompt after any significant build or change:
Run a full audit of this build. Check for: broken user flows, mobile layout issues, missing error states, Internet Identity edge cases, security gaps, and anything that would frustrate a real user.For a deeper audit focused on production readiness:
Run a production readiness audit. Focus on: canister upgrade safety, stable variable coverage, error handling for all async calls, input validation, and edge cases in the authentication flow.> How to Read Audit Reports
Audit reports come back with findings grouped by severity. Here's how to interpret them:
> How to Apply Fixes
After receiving an audit report, apply all fixes in a single targeted prompt:
Apply all the fixes from the audit report. Start with the critical issues, then warnings. Confirm each fix with a brief summary of what was changed.> Real Fix Examples
🔐 Internet Identity Edge Case
Found by audit: The app didn't handle the case where a user cancels the II login popup. The UI got stuck in a loading state with no recovery path.
Fix applied: Added timeout detection, error boundary, and a "Try again" button visible whenever the login flow is interrupted.
📱 Mobile Layout Issue
Found by audit: The data table overflowed horizontally on mobile, making key action buttons inaccessible on small screens.
Fix applied: Converted table to card layout on mobile with responsive breakpoints, all actions accessible at every viewport.
🏷️ Token Metadata
Found by audit: Token decimals were hardcoded to 8, causing display errors for tokens with different decimal precision.
Fix applied: Fetched decimals dynamically from the token canister and used them consistently throughout the UI.
// Audit Best Practices
- Run an audit after every major feature addition
- Always audit before sharing a build publicly
- Fix critical issues before moving to new features
- Re-audit after applying fixes — fixes can introduce new issues
- Use targeted audit prompts for specific concerns (mobile, security, etc.)
> Power Prompts — Take Control of Every Build
These 5 prompting patterns give you more control over how Caffeine builds your app. Use them at the start of a session or at any point when you want to slow down, review, or get structured feedback.
Think Before Building
Makes the AI explain its reasoning before writing any code — surfaces assumptions early.
Think step by step before building. Explain your approach first, then wait for my approval.Ask First, Build Second
Ensures the AI gathers the right context before starting — prevents costly misunderstandings.
Before you start building, ask me any clarifying questions you need to build this correctly.Explain Then Proceed
Builds accountability into your workflow — the AI must justify its approach before touching code.
Explain why this approach will work, then wait for my approval before proceeding.Step-by-Step Plan First
Lets you review the full plan before any code is written — ideal for complex features.
Tell me step by step what you're going to do, and wait for my approval before starting.Structured App Rating
Gets you actionable, structured feedback on your app concept — not vague opinions.
Rate my app 1–10 in three areas: idea, features, and user experience. Then suggest 3–5 specific things that would make it a 10/10 app.