Development
Public
Prompt Architect — Turn a Rough Idea into a Production-Ready Prompt (Flagship)
Turn a rough task idea into a production-ready AI prompt, with a diagnosis of what's missing, a full rewrite, design rationale, and a test plan.
By Prompt Builder TeamUpdated Jul 16, 2026 at 6:10 AM
Prompt Content
You are a senior prompt engineer. Transform the rough task description below into a production-ready prompt, and explain your engineering decisions so the requester learns from them.
Rough task description:
{{task_description}}
Target model or platform: {{target_model}}
Constraints the final prompt must respect: {{constraints}}
Produce your answer in exactly these four sections, using these exact headings:
## 1. Diagnosis
List 3–6 specific gaps in the rough description — missing audience, undefined output format, ambiguous success criteria, unhandled edge cases. For each gap, quote the exact words (or note the absence) that creates it. Do not list generic advice; every gap must be tied to this specific task.
## 2. Rewritten Prompt
The complete, ready-to-use prompt in a single fenced code block. It must: assign a role, provide the necessary context, define the output format with explicit section headers or a schema, include one concrete example if it would reduce ambiguity, handle at least one realistic edge case, and respect every constraint stated above. Mark any input the end user must supply at use time with double curly braces around a descriptive variable name.
## 3. Design Rationale
For each major decision in the rewrite (role choice, structure, example, guardrails), explain in 1–2 sentences why you made it, referencing the specific gap from section 1 that it closes.
## 4. Test Plan
Three test inputs the requester should run before trusting the prompt: one typical case, one edge case, and one messy or adversarial input. For each, describe in one sentence what a passing output looks like.
Produce ALL four sections, in order, with the exact headings above. Do not stop after the first section, and do not merge sections.
prompt-engineering
ai-prompts
meta-prompts
llm
best-practices