Code review, debugging, architecture, testing and docs. 100 reviewed prompts for engineering.
Diagnoses what makes working code hard to change and lays out small, low-risk refactoring steps with a safety net of tests and a before/after example. Use before cleaning up a messy module.
Reviews smart contract code for security flaws, risky admin powers and wasteful gas use, with exploit scenarios, fixes and tests to add. Use it before an external audit or any deployment that holds funds.
Critiques a product requirements draft or one-pager, ranking its weakest points, listing missing evidence and suggesting the questions to answer before building. Use before committing engineering time.
Reviews container orchestration manifests or deployment configs for reliability, security and cost problems and returns prioritized fixes with corrected snippets. Use before a production rollout or during a platform audit.
Audits a web page's code or test report for loading speed, accessibility, search basics and security hygiene, ranks issues by user impact and gives before/after fixes plus a pre-release checklist. Use it before launching or redesigning a page.
Designs a reliable data pipeline from named sources to an analytics-ready layer, covering layers, rerun safety, schema changes, quality checks, monitoring and backfills. Use when planning a new pipeline or reworking a fragile one.
Reviews how an application is organized and proposes a layered structure with small, testable steps. Use it when code has grown tangled and hard to change.
Recommends a practical build, test and deployment pipeline for your app, sized to your team and risk level, with a phased rollout. Use when setting up deployment for the first time or fixing slow, manual releases.
Answers small coding questions directly, and for bigger builds sets out the choices, design and folder structure first, then writes complete code in small, runnable steps. Use it to build a script, tool or small app with an assistant.
Explains a programming concept step by step with analogies, a small worked example, common mistakes and practice tasks. Use when mentoring a junior developer or learning a concept yourself.
Extracts what the feature is, what should happen and what actually happens from pasted logs, chat threads or rough notes, and flags missing details. Use before filing a ticket or handing a bug to a developer.
Designs how a machine learning model is trained, validated, deployed, monitored, and retired in production, including drift detection, rollback, and governance. Use when moving a model from notebook to production or fixing a fragile pipeline.
