Find the slow thing before optimising
This is slow: {{describe symptom, e.g. endpoint takes 4s at p95}}. Here's the code and what I know: {{paste code, query plans, timings, data volumes}}. Don't optimise yet. First: …
Coding & Debugging prompts
Practical coding & debugging prompts tested with useful AI apps.
Design review before you write the service
Review this design before I build it: {{paste design or describe: what it does, data model, endpoints, dependencies, expected scale}}. Attack it in this order: 1. FAILURE MODES — …
Review your own code before anyone else sees it
Here's the change I'm about to push: {{paste the diff}}. Context: {{what it's meant to do, what the codebase conventions are}}. Review it as three different people, separately: 1.…
Review this PR like the reviewer everyone respects
Review the diff below as a senior engineer on this codebase would. Order your review by what actually matters: 1. CORRECTNESS — logic errors, off-by-one, unhandled nulls, wrong er…
Write the docs the next engineer needs
Write documentation for the code below, for an engineer who joins in six months. Structure: 1. WHAT AND WHY — what this does and the problem it exists to solve. Include the constr…
Write the SQL, then explain what it will cost
Write the query for this question: {{describe what you need to know}}. Schema: {{paste DDL or table and column list with row counts}} Database: {{engine and version}} Requirement…
Turn a vague ticket into an implementation plan
Here's the ticket: {{paste}}. Here's the relevant code: {{paste}}. Produce an implementation plan: 1. Restate the requirement precisely, and list every ambiguity you had to resolv…
Understand an unfamiliar codebase fast
I've just been dropped into this codebase and need to ship a change to {{area}} this week. Given the files below, produce an orientation doc: 1. What this system does, in two sent…
Refactor without changing behaviour
Refactor the code below. The behaviour must not change, including its bugs, unless I approve the change. Process: 1. First, describe the current behaviour precisely, including edg…
Convert a script into something maintainable
This script works but it's about to become important: {{paste script}}. Harden it without over-engineering: 1. Name the failure modes it currently has — what happens on a network …
Turn a customer bug report into a reproducible test
Here's what the customer wrote: {{paste it verbatim, including the vague bits}}. Here's the relevant code: {{paste}}. Here's our test setup: {{framework, how fixtures work}}. Do t…
Unblock a dependency upgrade that broke the build
A dependency upgrade broke my build. Upgrade: {{package, from version, to version}}. Errors: {{paste the first and last 30 lines of the failure}}. Lockfile or manifest: {{paste rel…
Add logging to code you did not write
Here's a module I inherited and now have to support: {{paste the code}}. It currently logs {{nothing / everything at debug}}. Propose targeted logging, not a rewrite. For each sug…
Security pass on code that touches user data
Security-review the code below. It handles {{describe data and who can reach it}}. Check specifically: 1. AUTHZ — is every object access scoped to the caller's tenant and role, or…
Triage a test suite that fails at random
My CI is red often enough that people rerun jobs instead of reading them. Failure data: {{paste test names, failure counts, and recent failure output}}. Stack and test runner: {{e.…
Plan a schema migration that runs under live traffic
I need to change a database schema while the app keeps serving traffic. Current shape: {{tables, columns, types, indexes, row counts}}. Target shape: {{what it should become}}. Sta…
Estimate work you have never done before
I've been asked how long this will take: {{describe the work}}. What I know: {{stack, team size, what already exists}}. What I've never done before: {{the unfamiliar part}}. Break…
Decide the error handling for a new API endpoint
I'm adding an endpoint: {{method, path, what it does}}. It touches {{database / third-party API / queue}}. Callers are {{our own frontend / third-party integrators / both}}. Give …
Draft the postmortem from a raw incident timeline
Write a blameless postmortem from the raw material below. Incident channel or timeline: {{paste messages, alerts, and deploy events with timestamps}}. User impact as measured: {{wh…
Debug the thing that only fails in production
This works locally and fails in production. Symptoms: {{describe, including frequency and pattern}}. Environment differences I know of: {{list}}. Logs: {{paste}}. Work the problem…
Explain this error like a senior engineer would
Here's an error and the relevant code. Diagnose it properly. 1. Translate the error into plain language: what the runtime was trying to do when it gave up. 2. Name the actual caus…
Write the tests that would have caught the bug
Here's a bug that reached production, and the code around it. 1. Write the failing test that reproduces it at the smallest useful level (unit if the logic is pure, integration if …
Write the on-call runbook for the service you own
You are a staff engineer writing an on-call runbook for someone who has never touched this service. The service: {{what it does, who depends on it, what it talks to}}. Paste the co…
Write the PR description reviewers stop asking questions about
Here's my diff: {{paste git diff or the file list plus key hunks}}. The ticket said: {{paste it}}. Write the pull request description with these sections: 1. WHAT CHANGED — two se…
Decide whether to roll back or fix forward at 2am
You are the calm senior engineer on an incident bridge. Symptoms and blast radius: {{errors, affected users, when it started}}. What shipped recently: {{commits, migrations, config…
Decide what to delete from a codebase nobody has cleaned in two years
You are a staff engineer doing a deletion pass. Candidate dead code: {{paste file list, functions, feature flags, endpoints, and any usage or analytics data you have}}. How we dete…
Split one giant change into pull requests someone will review
You are a senior engineer breaking an oversized change into a review-friendly series. Below is the diff summary (or file list plus description) and the constraints. Produce: 1. PR…
Work out why your service is slow for one customer only
You are a senior engineer debugging a latency complaint that only one customer sees. Below are their symptoms, our timings and the relevant code or queries. Work the problem. 1. S…
Cut a twenty-minute build down to something you can live with
You are a build engineer looking at a CI pipeline that has crept up to {{current wall-clock time}}. Config: {{paste CI config / Dockerfile / build script}}. Timing breakdown per st…
Decide what to do with the feature flags nobody ever cleaned up
You are a staff engineer doing a flag cleanup on a codebase where every flag was supposed to be temporary. Flag list with creation dates, current rollout state, and owner if known…
Decide what to do with a pull request that has been open for three weeks
You are a staff engineer triaging a stale pull request. It is too big, the branch has drifted, and nobody wants to be the one to review it. The PR description and diff summary (fi…
Decide what to do about a bug you cannot reproduce
You are a senior engineer on call. A bug is reported but does not reproduce locally. Report: {{paste the ticket or user description}}. Relevant code: {{paste}}. What we have from p…
Decide how your code should behave when a vendor API goes down
You are a staff engineer reviewing a failure path. Below is the code that calls a third-party API, plus what the user is doing when it runs and what the vendor's failure modes look…
Write the deprecation plan for an endpoint customers still call
You are the engineer who owns an API endpoint that has to go. Below are the endpoint and its replacement, request volume by client or API key over the last month, and why we are re…
Review a dependency before you add it to the codebase
You are a staff engineer who has been burned by abandoned libraries. I am about to add a dependency. Judge it in five sections, no preamble. 1. WHAT IT ACTUALLY DOES - two sentenc…
Write the load test that mirrors how your service is actually used
You are a performance engineer. Below is my endpoint list with real traffic shape and the infrastructure behind it. Design the load test. 1. THE PROFILE WORTH TESTING - table: End…
Cut your cloud bill without breaking the service
You are a pragmatic infrastructure engineer. I need to cut spend without a reliability incident that costs more than the savings. Cost breakdown by service or line item, last full…
Decide whether to rewrite a module or keep patching it
You are a staff engineer who has seen rewrites fail. Below is a module the team keeps patching, its recent bug history and what we need from it next. Tell me whether to rewrite it,…
Clean up the bad data a bug wrote into production
You are the engineer who has done this before and knows a rushed backfill causes the second incident. A bug wrote incorrect rows to production. Below are the schema, the bug, and t…
Find the change that broke a feature that worked last release
You are a senior engineer who finds regressions by narrowing, not by reading every diff. A feature worked in the previous release and is broken now. Help me find the commit. Return…
Contain an API key someone just committed to the repo
You are a security-minded senior engineer. A credential was committed to our repository. Below is what leaked, where, and for how long. Walk me through containing it in the right o…
Close the gap between your test environment and production
You are a senior engineer on a team where tests pass in staging and break in production. Below is how each environment is configured and the last few incidents that got through. Pr…
Track down a memory leak in a service that slowly falls over
You are a senior engineer who has chased memory leaks in production. My service's memory climbs until it restarts. Using the evidence below, help me find the cause in five sections…