Coding & Technology prompt builder
Say what you want in plain words, pick your AI, and get a clear, structured prompt. Everything runs in your browser. No sign-up or API key needed.
Your prompt
For ChatGPTReview my React component for bugs and performance · Coding › Code reviews
Regenerate cycles through alternative structures for the same inputs. These actions are rule-based and instant: no AI API is called, and your inputs are kept when you switch platforms.
## Role You are a staff engineer doing a thorough code review focused on correctness, security, and maintainability, in that order. ## Task Review my React component for bugs and performance. Success looks like: I get prioritised, actionable findings with fixes, so I know exactly what must change before merging. ## If details are missing I haven't told you these yet, so assume: - Code or diff: I'll paste it next - Stack: infer from the code - Purpose: infer from the code and state it State these assumptions in one line at the top of your answer. If one would change the answer a lot, ask me about it first. ## Steps 1. Summarise what the code does in 2 lines 2. Check correctness and edge cases 3. Check security (injection, auth, secrets, input validation) 4. Check performance and maintainability 5. Give fixes for every high-severity finding ## Constraints - Group findings by severity: blocker, should fix, nit - Quote the exact line for each finding - Don't nitpick style a linter would catch unless asked - No filler: skip intros, restating my request and closing summaries - Don't invent facts, statistics, quotes, sources or details about me; flag anything I should verify ## Output format - **Summary** - **Findings** table: severity | line | issue | fix - **Fixed code** for blockers - **Tests to add** ## Before you answer, check - Every blocker has a concrete fix - Findings reference real lines
Copies your prompt and opens the official site in a new tab. You may need to paste it yourself (Ctrl/⌘ + V). Nothing is sent to the AI by us.
Export and standing instructions
SKILL.md is an Agent Skill file (name + description frontmatter) for Claude and other agents that support skills.
Paste into a new ChatGPT chat, or into a custom GPT's instructions.
## Role You are a staff engineer doing a thorough code review focused on correctness, security, and maintainability, in that order. ## Task Review my React component for bugs and performance. Success looks like: I get prioritised, actionable findings with fixes, so I know exactly what must change before merging. ## If details are missing I haven't told you these yet, so assume: - Code or diff: I'll paste it next - Stack: infer from the code - Purpose: infer from the code and state it State these assumptions in one line at the top of your answer. If one would change the answer a lot, ask me about it first. ## Steps 1. Summarise what the code does in 2 lines 2. Check correctness and edge cases 3. Check security (injection, auth, secrets, input validation) 4. Check performance and maintainability 5. Give fixes for every high-severity finding ## Constraints - Group findings by severity: blocker, should fix, nit - Quote the exact line for each finding - Don't nitpick style a linter would catch unless asked - No filler: skip intros, restating my request and closing summaries - Don't invent facts, statistics, quotes, sources or details about me; flag anything I should verify ## Output format - **Summary** - **Findings** table: severity | line | issue | fix - **Fixed code** for blockers - **Tests to add** ## Before you answer, check - Every blocker has a concrete fix - Findings reference real lines
Before vs after
Before · what you typed
Review my React component for bugs and performance
8 words
After · your prompt
## Role You are a staff engineer doing a thorough code review focused on correctness, security, and maintainability, in that order. ## Task Review my React component for bugs and performance. Success looks like: I get prioritised, actionable findings with fixes, so I know exactly what must change before merging. ## If details are missing I haven't told you these yet, so assume: - Code or diff: I'll paste it next
245 words · full prompt above
What we added
- Expert role: staff engineer doing a thorough code review focused on correctness, security, and maintainability, in that order.
- Success criterion: what a great answer must achieve
- Stated assumptions: sensible defaults you can correct
- Step-by-step workflow: 5 steps for this task type
- Constraints: 5 rules (length, clichés, no invented facts)
- Output format with tables: 4 sections defined
- Self-check: the AI reviews its answer before replying
Make it more specific
Answering any of these improves the result. Or press “Add missing context” to insert [placeholders] and have the AI ask you.
- ○ Please paste the code or diff.
- ○ Language, framework and versions?
- ○ What is this code supposed to do?
- ○ What format and length do you want?
Prompt checklist
25/100A completeness checklist, not a prediction of output quality.
- Clear goal
- Context: Add background the AI can't guess.
- Audience: Who is the result for?
- Output format: Name the format and length (a sensible default is used otherwise).
- Tone or constraints: Add a tone or at least one rule.
- Example / reference: An example of good output improves consistency.
- Task details: Fill in the Code reviews details (e.g. Code or diff).
Prefer fixed templates? Try the classic template builder.
Platform guides: ChatGPT · Claude · Gemini · Grok · Other AI assistant · Perplexity · Cursor / coding agents · Claude Code · Midjourney · Image · Video