780 reviewed prompts.
Produces a design-first plan to improve hierarchy, clarity, trust and conversion of an existing page. Use it before touching the page so changes are agreed first.
Reviews an AI agent or automation design for human control, handling of unclear instructions, prompt-injection defenses, transparency and data exposure, and names concrete fixes. Use it before giving an agent real tools or data.
Diagnoses a slow endpoint, job or page from your symptoms and measurements, and returns likely root causes, a ranked fix list with expected gains and a before/after measurement plan. Use it before scaling hardware.
Maps how a feature or system you own could be attacked or misused and lists prioritized defenses, detection and tests. Use during design or before a security review of your own product.
Goes through a hospital bill and insurance policy text to list items an insurer commonly excludes, with totals and questions to raise. Use before filing a claim or disputing a bill.
Checks an A/B test for setup problems, interprets the lift and its uncertainty, reviews guardrail metrics, and gives a clear ship, iterate or stop recommendation. Use when an experiment has finished and someone has to decide.
Reviews an infrastructure-as-code change and its plan output for destroyed or replaced resources, leaked secrets, blast radius and version drift, then gives a validation and rollback plan. Use it before applying changes to shared environments.
Standing approach for reviewing code: rank findings by impact, explain why, and suggest a fix. Use when asked to review code, a change or a pull request.
An interview that helps you turn a fuzzy aim into a clear goal, limits and success measures, with no advice until you are ready. Use it before planning a project, career move or change.
Reviews a Python snippet for missing docs, style, type hints and slow patterns, then returns an improved version with a change list. Use it before sharing or merging a script.
Pulls the fields you define out of messy text (emails, notes, documents) and returns strict, valid JSON with consistent types and nulls for missing values. Use it to turn unstructured text into rows for a spreadsheet or system.
Condenses a long conversation into decisions, instructions, open questions and next steps so you can continue it later or share it. Use when a chat gets long or you switch tools or people.
