Loading tool…
prompts
Here's the error I'm getting: [paste error]. Here's the relevant code: [paste code]. Explain what's causing it and give me the minimal fix.
Root-cause a specific error without a full rewrite
This function should do X but instead does Y. Walk through it step by step and tell me where the logic diverges from what I expect.
Trace logic bugs when behavior doesn't match intent
Add temporary debug logging at the key decision points in this function so I can see what's happening at runtime.
Instrument code to gather more info before fixing
This bug only happens intermittently / in production. List the most likely causes given this code and stack.
Narrow down flaky or environment-specific bugs
Here's my code and a failing test. Explain why the test fails, then fix the code, not the test.
Keep the test as the source of truth while fixing implementation
Bisect this: here's the working version and the broken version. What changed that could cause this regression?
Diff-driven debugging for regressions after a change
Refactor this function for readability without changing its behavior. Explain each change you make and why.
Safe refactor with rationale, not a silent rewrite
This file has grown too large. Suggest how to split it into smaller modules, and show the new file boundaries.
Break up god-files/modules along sensible seams
Point out any code smells, duplication, or anti-patterns in this file, ranked by how much they'd hurt maintainability.
Prioritized quality review instead of a generic list
Rewrite this using [pattern/library, e.g. early returns / a hook / composition] instead of the current approach, and explain the tradeoff.
Targeted pattern-swap refactor with explicit tradeoffs
Simplify this without adding new abstractions or dependencies — I want fewer lines and fewer concepts, not more indirection.
Guards against over-engineered 'refactors'
Is this function doing too many things? If so, split it and name each piece by what it actually does.
Single-responsibility check with concrete splitting
Write unit tests for this function, covering the happy path, edge cases, and at least one failure case.
Baseline test coverage prompt for a single function
Here's a bug that just got fixed: [describe/paste diff]. Write a regression test that would have caught it.
Turn a fixed bug into a permanent regression test
What are the edge cases I'm not testing for this input? List them, then write tests for the ones that matter most.
Surface missed edge cases before writing tests
Write an end-to-end test for this user flow: [describe flow, e.g. sign up -> verify email -> log in].
Scaffold an e2e test from a plain-English flow description
This test is flaky. Explain the likely source of non-determinism and rewrite it to be stable.
Fix flaky tests instead of just re-running them
Mock out [external dependency, e.g. the API call / database] in this test so it runs in isolation.
Isolate unit tests from external dependencies
Review this PR diff as if you were a senior engineer. Flag bugs, security issues, and unclear naming — skip style nitpicks.
Substance-focused review, filters out noise
Is there anything in this change that could break existing callers or break backwards compatibility?
Check for breaking changes before merging
Review this for security issues specifically: injection, auth, secrets, unvalidated input.
Narrow, security-only review pass
Summarize what this PR does in 2-3 sentences for the changelog, based on the diff.
Generate a plain-English PR/changelog summary
Would a new engineer understand this code without extra context? Point out anything that needs a comment or better naming.
Readability check from a newcomer's perspective
Compare these two implementations of the same feature and recommend which to merge, with reasons.
Decide between competing implementations objectively
Write a docstring/comment for this function describing its purpose, parameters, return value, and any side effects.
Standard function-level documentation
Generate a README section explaining how to set up and run this project locally, based on the package.json/config files.
Bootstrap a setup guide from existing config
Add inline comments only where the code isn't self-explanatory — don't narrate the obvious lines.
Avoid comment noise while covering the tricky parts
Document this API endpoint: method, path, request body, response shape, and possible error codes.
Quick API reference doc from route code
Explain this codebase's architecture in a short doc for a new team member — main modules and how data flows between them.
Onboarding-style architecture overview
Turn these commit messages / PR titles into a human-readable changelog entry.
Convert raw git history into release notes
I want to build [feature]. Before writing code, outline the approach: main components, data flow, and key decisions.
Force a plan before implementation begins
Given these requirements: [list], what are 2-3 possible architectures, and what are the tradeoffs of each?
Compare architectural options before committing
How would this design scale if usage grew 10x? What would break first?
Stress-test a design against future scale
Break this feature down into small, independently shippable tasks I can work through one at a time.
Decompose a large feature into a task list
What existing patterns/modules in this codebase should this new feature reuse instead of duplicating?
Encourage reuse over reinvention within a codebase
Review this database schema for this use case: [describe]. Point out normalization issues or missing indexes.
Sanity-check a schema before it's hard to change
Give me the relevant code only, not the whole file, and tell me exactly which lines to replace.
Cuts down on regenerated boilerplate you didn't ask for
Before making changes, ask me any clarifying questions you need — don't assume.
Reduces wrong-guess implementations on ambiguous asks
Explain your reasoning first, then show the code — I want to check the approach before the implementation.
Catches bad approaches before code gets written
Don't add new dependencies, frameworks, or files unless I explicitly ask for them.
Keeps AI-driven changes minimal and in-scope
Here's my code style/conventions: [describe or paste example]. Match this style in your suggestions.
Keeps AI output consistent with existing codebase style
Give me three different ways to solve this, ranked by simplicity, and explain when I'd pick each one.
Surfaces alternatives instead of one default solution
Get new cheatsheets & tools in your inbox
No spam — just new releases, occasionally.