Your build starts here
Replace bracketed details with your own goal, context, and checks. Keep the first build small enough to inspect.
Inspect this existing prototype: [project or repository]. Its intended audience is [users], and its main journey is [specific action through confirmed result]. First run it using its documented setup and identify the files and services involved in that journey.
Create a small inventory with status observed working, observed failing, or not checked. For each item, record the trigger, expected behavior, actual behavior, and the evidence you used. Include refresh or relaunch, an empty state, one invalid input, and one failure of a dependency when they apply. Use disposable fixtures and describe any check you cannot perform.
Then choose the smallest useful gap that blocks the intended journey. Explain why it matters and propose a bounded change with acceptance cases. Do not substitute a redesign or a stack migration for this inspection. If I authorize implementation, reuse the current conventions and finish by showing the changed behavior, relevant diff, and checks actually performed.Make this prompt yours
Use when a demo looks promising but you cannot tell which features are durable. Give the assistant a runnable project or enough setup information; a screenshot alone cannot establish stored behavior.
You might also hear
- vibe coding audit
- feature inventory
- unfinished app
The docs behind it
These references explain the underlying concepts and APIs. The prompt templates and project plans are original examples to adapt to your own setup.
Research sources reviewed .
Editorial review · October 28, 2026
Supported: Original, unexecuted inspection template. It asks for observed evidence and unchecked states rather than inferring completion from a feature list; Anthropic's testing failure modes and GitHub review limits support that caution.