Coding prompts

Debugging, reviews, refactors and tests. Pick one and open it in Promptea.

Debugging with reproducible steps

I’m hitting this bug:

Context:
- Stack: [Next.js / Node / Python / etc.]
- Environment: [local/prod]
- Expected:
- Actual:

Logs:
[paste logs]

Minimal code:
[paste snippet]

I need:
1) most likely hypothesis
2) repro steps
3) fix (diff/snippet)
4) edge cases
If missing info, ask 1–3 questions first.

Opens on home with the prompt prefilled.

Open in Promptea

Code review with criteria

Review this code:

[paste code]

Evaluate:
- Correctness
- Security
- Performance
- Readability
- Missing tests

Return:
- prioritized issues
- concrete fixes (snippets)
- test checklist

Opens on home with the prompt prefilled.

Open in Promptea

Feature design (step plan)

I want to implement: [feature]
Constraints: [time, stack, limits]
Users: [who uses it]

Propose:
- simple architecture
- endpoints/modules
- data model
- edge cases
- step plan

Opens on home with the prompt prefilled.

Open in Promptea

Refactor with clear goals

Refactor this code with goals:
- reduce complexity
- improve naming
- separate responsibilities
- keep behavior (don’t break API)

Code:
[paste code]

Return:
- brief explanation
- diff/snippets
- suggested tests

Opens on home with the prompt prefilled.

Open in Promptea

Write tests (unit + integration)

I need tests for this module:

Stack: [jest/vitest/etc.]
Code:
[paste code]

Provide:
- cases to cover (list)
- unit tests (snippets)
- 1–2 integration tests
- how to run them

Opens on home with the prompt prefilled.

Open in Promptea

Performance: diagnose + optimize

I have a performance issue.

Context:
- app: [web/api]
- symptom: [latency/CPU/memory]
- metrics: [p95, RPS, etc.]

Code/logs:
[paste]

Give me:
- hypotheses by impact
- how to measure (profiling)
- fixes and tradeoffs

Opens on home with the prompt prefilled.

Open in Promptea

FAQ

How do I use these templates?

Pick an example and click “Open in Promptea”. It pre-fills the prompt, purpose (Code), and a suggested model.

What should I paste (minimum)?

A minimal snippet, relevant logs, and expected vs actual behavior. That improves score and output quality.

How do I avoid generic answers?

Request an output format (steps, diff, checklist) and add constraints (stack, versions, limits).

Can I change the model?

Yes. For strong structure, try Claude/GPT. For quick bullets, Grok can work well.