Copy · paste · learn
Prompt library
50 of 105 prompts
Agent Task Planner
It separates judgment from fixed steps, which keeps agents cheap and predictable.
agents
BuilderDeepSeekMembersAlgorithm Reasoning Walkthrough
Tracing a small example by hand is a practical way to catch logic errors in reasoning-focused models.
coding
BuilderAny assistantMembersBackup and Restore Plan Builder
Starting from recovery targets rather than tools produces a plan that fits the business.
it-ops
BuilderAny assistantMembersBash Script Safety Review
It asks about the specific classic shell hazards, so the review is concrete rather than a vague "looks fine".
it-ops
BuilderAny assistantMembersBuild a Regular Expression with Test Cases
Positive and negative examples force precision.
coding
BuilderAny assistantMembersChange Request Summarizer
It mirrors what approvers actually look for, so the request gets fewer follow-up questions.
it-ops
BuilderPerplexityMembersCited Answer from a Search Assistant
Specific scope focuses the search and the source preferences push it toward better pages.
research
BuilderClaudeMembersClaude Project Instructions Starter
Project instructions work best when they explain how to treat uploaded material, including what to do when the answer is missing.
productivity
BuilderClaudeMembersCodebase Onboarding Guide
Tracing one concrete flow teaches architecture better than a list of folders.
coding
BuilderDeepSeekMembersComplexity and Bug Check
Requiring a triggering input for each bug filters out false alarms, which are a common problem in automated reviews.
coding
BuilderClaudeMembersContract Summary for Non-Lawyers
Asking for section references lets you verify each point against the source, which is the main safeguard with long documents.
analysis
BuilderGrokMembersCross-Check a Trending Claim with Grok
Separating confirmed from unconfirmed fits fast-moving stories where early posts are often wrong.
research
BuilderAny assistantMembersDebug an Error Step by Step
It supplies the same facts a human helper would ask for.
coding
BuilderAny assistantMembersDebug My Prompt
Showing the gap between wanted and received output gives the model a concrete target.
productivity
BuilderAny assistantMembersDecision Pre-Mortem
Assuming failure has already happened frees the model to name risks it would otherwise soften.
business
BuilderAny assistantMembersDefensive Security Review Checklist
Declaring scope and authorization up front keeps the output defensive and focused.
security
BuilderChatGPTMembersDraft Instructions for a Custom GPT
The sections mirror what makes custom assistants reliable.
agents
BuilderGeminiMembersEquipment Inventory From Photos
A confidence column and an explicit unreadable marker keep the inventory honest.
it-ops
BuilderGeminiMembersExplain a Diagram Image
Listing components and connections explicitly lets you check what the model saw before you rely on its explanation.
it-ops
BuilderAny assistantMembersExtract Fields to JSON
Naming allowed values and the null rule removes the usual ambiguities.
coding
BuilderAny assistantMembersFact-Check Checklist for My Draft
It uses the model to find what to check, not to certify truth, which respects its weaknesses.
research
BuilderAny assistantMembersFind the Weak Points in My Argument
Assigning the role of a skeptic overrides the model's tendency to agree.
analysis
BuilderAny assistantMembersFirst-Pass Data Analysis from a Table
It starts with understanding and cleaning, which is where most analysis problems begin.
analysis
BuilderAny assistantMembersGenerate Unit Tests for a Function
Naming the categories of cases pushes beyond happy-path tests.
coding
BuilderChatGPTMembersHelpdesk Assistant Custom GPT Instructions
Spelling out what the GPT must not do and when it escalates is what makes a support bot safe to use.
it-ops
BuilderAny assistantMembersIncident Post-Incident Report Writer
The [ASSUMED] tag separates facts from filler, which is the biggest risk when a model writes an incident report.
it-ops
BuilderLlamaMembersLog Line Classifier
A closed label set and a fixed output format suit local models and make results easy to scan or paste into a spreadsheet.
it-ops
BuilderAny assistantMembersLog Triage Assistant
It separates pattern-finding from guessing by forcing the answer to cite specific lines.
it-ops
BuilderAny assistantMembersMatch My Writing Voice
Real samples teach style better than adjectives like "casual".
writing
BuilderChatGPTMembersMessy Spreadsheet Cleanup With a Script
A change log and flagged rows mean you can audit the cleaning rather than trust it.
data
BuilderAny assistantMembersNegotiation Preparation Partner
Interests and a walk-away point move preparation from positions to trades.
business
BuilderAny assistantMembersNetwork Description Explainer
The ASCII diagram lets you verify the model understood your network before you trust its advice.
it-ops
BuilderMicrosoft CopilotMembersPolicy Draft in Word
Numbered, testable rules make a policy enforceable and easier to audit.
writing
BuilderAny assistantMembersPowerShell Script Review
Naming where and how the script runs lets the model check version and permission problems instead of giving generic advice.
it-ops
BuilderAny assistantMembersPre-Merge Code Review
A defined checklist and table make the review consistent and scannable.
coding
BuilderGrokMembersPublic Reaction Summary
Treating social reaction as a biased sample, not a poll, prevents overconfident conclusions.
analysis
BuilderClaudeMembersRefactor for Readability Without Changing Behavior
The behavior-preservation constraint prevents hidden rewrites.
coding
BuilderGrokMembersRumor Check With Caveats
A fixed verdict scale with a "cannot determine" option stops the model from forcing certainty.
research
BuilderPerplexityMembersSecurity Advisory Lookup
Pointing the search at official advisories first reduces reliance on rumor.
security
BuilderMistralMembersShort JSON Extraction
Defining fields, types and the null rule gives a predictable output that other programs can parse.
data
BuilderAny assistantMembersShort Proposal Writer
Anchoring the proposal in the client's own words keeps it about them rather than you.
business
BuilderAny assistantMembersStress-Test a Business Idea
Focusing on assumptions and cheap tests turns a vague plan into experiments.
business
BuilderClaudeMembersStructured Prompt With Tagged Sections
Tagged sections keep different kinds of information clearly separated, which helps most assistants avoid mixing instructions with source text.
productivity
BuilderAny assistantMembersStructured Research Brief
The structure demands disagreement and open questions, not just a tidy summary.
research
BuilderAny assistantMembersSystem Prompt Generator
Clarifying questions surface missing requirements before drafting.
agents
BuilderAny assistantMembersTechnical Interview Prep Coach
One question at a time makes it feel like a real interview and gives you time to think.
business
BuilderLlamaMembersTest a Local Open Model Fairly
Five small tasks that you can judge yourself reveal summary quality, format discipline, coding, explanation and reasoning.
learning
BuilderClaudeMembersTough but Fair Editor Feedback
A fixed order of feedback stops a vague "looks good".
writing
BuilderPerplexityMembersVendor Comparison With Sources
Requiring a citation for every cell lets you click through and verify the claim instead of trusting a summary.
research
BuilderDeepSeekMembersWork Through a Problem and Verify It
A second method catches slips that a single pass can hide.
analysis