The Lazy Loader
Bypasses manual editing of prompt variables. It forces the AI to interview you to establish the "Rules of Engagement" (Role, Tone, Format) before it attempts any work.
Tags: Setup, Configuration, Utility
Free, production-ready AI prompts and agent skills from Kyt Services — the same ones we use on engagements, open-sourced CC-BY 4.0. 69 prompts across 13 categories.
Bypasses manual editing of prompt variables. It forces the AI to interview you to establish the "Rules of Engagement" (Role, Tone, Format) before it attempts any work.
Tags: Setup, Configuration, Utility
Solves the "Context Window Limit." When the AI starts forgetting things or gets slow, this packs the entire session's logic into a dense "Save File" that you can load into a new chat. Think of this as a "Save Game" point for long projects.
Tags: Context, Migration, Utility
Stops the AI from trying to be helpful too early. When you upload a big PDF, the AI usually summarizes it immediately—wasting tokens and missing details. This command forces it to "Read Only" and wait for instructions.
Tags: Documents, Context Loading, Analysis
This is the "BIOS Update" for the model. It overrides the default "helpful assistant" persona (which tends to be sycophantic) with a "Critical Partner" persona that prioritizes truth over politeness.
Tags: System, Red Team, Critical Thinking
This acts as a "Chain of Thought" enforcer. By forcing the AI to list its assumptions and checks before answering, you significantly reduce hallucination rates. It also employs a mandatory Clarification Gate to prevent incomplete outputs.
Tags: Chain of Thought, QA, Verification
AI models suffer from "context drift" and often become yes-men to please the user. This prompt forces a hard reset on the logic engine, compelling it to attack your ideas rather than validate them.
Tags: Red Team, Correction, Critical Thinking
Converts the "morning chaos" of unread messages into a structured battle plan. It specifically looks for "Soft Deadlines" (e.g., "Need this by Friday") that are buried in casual text.
Tags: Email, Prioritisation, Eisenhower Matrix
Similar to the Daily Driver but optimised for Tasks rather than Status. It excels at identifying "Fake Urgency" (tasks that are noisy but not important) and moving them to P3/P4.
Tags: Backlog, Prioritisation, Eisenhower
The "Forensics of Work." It takes a messy thread of "Reply All" emails and extracts the only thing that matters: The Decision. It ignores the fluff and highlights the blockers.
Tags: Email, Synthesis, Executive
The ultimate time-saver. It separates "Chatter" (what was discussed) from "Decisions" (what is locked in). The Action Table enforces accountability by requiring an "Owner" and a "Date."
Tags: Meeting, Notes, Actions
Detects "Watermelons" (Green on the outside, Red on the inside). It reads the commentary text to see if it contradicts the RAG status (e.g., Status: Green, but Comment: "We are currently blocked by legal").
Tags: Status, Watermelon, RAG
Acts as a senior PMO Director scrutinizing a newly pitched initiative. It forces strict evidence tagging ([EXPLICIT] vs [INFERRED]) to stop the AI from hallucinating project scope, deliverables, or timelines that don't actually exist in the raw documents.
Tags: Project, Scoping, Audit
Forces proper Risk Syntax (Cause -> Risk -> Effect). It prevents the common error of listing "Staffing" (a category) as a risk, ensuring you write "Staffing shortage leads to missed deadline."
Tags: Risk, Register, Mitigation
Many PMOs confuse project delivery (on time/on budget) with value realisation. This prompt acts as a ruthless auditor. It compares the original approved business case against current operational metrics to detect "benefit shortfalls" early, proposing strategic pivots if the ROI is failing.
Tags: Benefits, ROI, Value
Moves beyond individual task triage. It identifies systemic single points of failure across the portfolio (e.g., relying on the same senior engineer for three critical-path projects simultaneously) and suggests scenario plans to alleviate burnout.
Tags: Resources, Capacity, Bottlenecks
As PMOs deploy Agentic AI systems that can execute multi-step workflows across systems, the risk of unauthorized actions, data misuse, and systemic disruptions rises significantly. This prompt stress-tests the workflow to ensure irreversible actions are blocked without explicit human authorization.
Tags: AI, Governance, Security
An honest opportunity-map for PMO leaders: scores candidate workflows on volume, reversibility, data, failure cost, and integration. Names what to defer, what to never automate, and a 90-day pilot for the top candidate.
Tags: PMO, Agentic AI, Operating Model
A RAG (Retrieval Augmented Generation) controller. It prevents the common AI error of treating a "Draft" document as fact. It forces the AI to check headers for "Version 0.1" vs "Final".
Tags: RAG, Documents, Summary
When you need to know "Who knew what, and when?", this prompt acts as a digital detective. It forces the AI to search for synonyms and verify file attachments before building a timeline.
Tags: Timeline, Forensics, Investigation
Moves beyond Wikipedia-style summaries. It forces the AI to analyse the mechanics of a topic and identify the "Red Team" counter-arguments. Crucial for preparing for high-stakes meetings where you might be challenged on the details.
Tags: Analysis, First Principles, Research
A dual-mode tool. Mode A generates complex Excel/DAX syntax (correcting for common syntax errors). Mode B analyzes raw data tables to find outliers/trends you might miss with the naked eye.
Tags: Excel, Data, Statistics
Designed for AI tools that are connected to your file system (e.g., Copilot M365). It adds a layer of logic to the search to prioritise "Final" over "Draft" and recent over old.
Tags: SharePoint, Search, Documents
Avoids the "This isn't what I asked for" loop. By agreeing on the Table of Contents (Outline) first, you ensure the AI writes the right content in Step 3.
Tags: Document, Structure, Planning
Most organisational calendars are "additive" (people add meetings but never remove them). This prompt forces a "Zero-Based" approach: it assumes no meetings exist and forces you to justify every single one back into existence based on pain points.
Tags: Meetings, Rhythm, Organisation
This is a "Logic Auditor" for your ideas. It ensures you have a "Hero" (the user/customer) and a "Conflict" (the problem) before you start writing slides.
Tags: Narrative, Pitch, Strategy
Prevents "Announcement Regret." Most change fails because the "After" state is vague or the "Bridge" (resources) is missing. This prompt forces a "Pre-Flight Check" to see if you are actually ready to launch.
Tags: Change, Strategy, Communication
A start-to-finish toolkit for facilitators. Mode A plans the logistics (including the pre-work email). Mode B cleans up the mess of post-it notes afterwards into themes.
Tags: Workshop, Facilitation, Design
Acts as a seasoned CFO evaluating corporate health. Rather than providing generic advice, it forces the AI to pinpoint specific operational bottlenecks and propose targeted fixes.
Tags: Finance, Profit, Operations
Use this when an organisation is experiencing friction between its strategy and execution. It systematically audits the hard and soft elements of the business (Strategy, Structure, Systems, Shared Values, Style, Staff, and Skills) to ensure cross-functional alignment.
Tags: McKinsey, Organisation, Audit
Helps identify external threats and opportunities before a major product launch or market entry. It enforces the Pyramid Principle, ensuring the AI delivers the "Bottom Line Up Front" before diving into the supporting evidence.
Tags: PESTLE, Strategy, Framework
Serves as a ruthless "Mock VC." It evaluates early-stage or expansion pitches, identifying the weakest links in the narrative before you put the deck in front of actual investors.
Tags: Pitch, VC, Investment
The foundational tool of Design Thinking. It turns a "dead end" complaint into a "solvable" question.
Tags: Design Thinking, Ideation, HMW
A systematic checklist to force iteration. If you have a "Version 1.0" idea, this prompt forces you to consider "Version 2.0" by modifying specific variables (Substitute, Combine, Eliminate, etc.).
Tags: SCAMPER, Innovation, Iteration
Breaks "Functional Fixedness" by imposing extreme, artificial constraints (e.g., "Do it for free" or "Do it in 60 seconds").
Tags: Brainstorming, Constraints, Creativity
This is a "Reference Card" for you, not a prompt for the AI. It contains the best questions to ask a room full of people to draw out creativity.
Tags: Facilitation, Cheat Sheet, Workshop
Writes emails that people actually read. It enforces "BLUF" (Bottom Line Up Front) and "Action-Oriented" subject lines. It strips out the "I hope this email finds you well" fluff.
Tags: Email, Communication, BLUF
Your "Brand Police." It enforces New Zealand English, correct Te Reo Māori usage, and cuts out "corporate fluff." Good for Teams Channels, Intranet Posts, and Newsletters.
Tags: Comms, Brand, Writing
Solves the "Frankenstein Deck" problem where every slide looks different. Mode A "learns" your company style. Mode C actually writes Python code that you can run to create the .pptx file automatically.
Tags: PowerPoint, Slides, Presentation
Fixes the flow of a presentation. It calculates timing (approx 90s per slide) to ensure you don't end up with a 50-slide deck for a 30-minute meeting.
Tags: Presentation, Structure, Timing
Analogies are dangerous if they break down under scrutiny. This prompt generates the analogy and explicitly tells you where it stops working (the "Leak"), so you don't look foolish when questioned.
Tags: Analogy, Communication, Explanation
The "COPE" strategy (Create Once, Publish Everywhere). It takes a core asset (like a whitepaper) and atomizes it into social posts, newsletters, and memos.
Tags: Content, Repurpose, Marketing
Removes the "Halo Effect" (liking someone so ignoring their faults) or "Recency Bias" (only remembering the last 2 weeks). It acts as a neutral "Performance Coach" that forces you to prove your claims with evidence.
Tags: HR, Performance, Review
Shifts your CV from "Duties" (what you did) to "Results" (what you achieved). It uses a Gap Analysis to find exactly which keywords from the Job Description are missing from your profile and injects them naturally.
Tags: CV, Career, Cover Letter
Generates the highly technical "Camera Settings" text needed to make an AI image look like a photograph (e.g., 85mm lens, f/2.8 aperture, Rembrandt lighting).
Tags: Image Gen, Headshot, Creative
A "Prompt Engineer in a box." It doesn't just write the instructions; it interviews you to find out what you really need and builds "Failure State" handling (what the bot should do when it gets confused).
Tags: AI Agent, Custom GPT, Design
Acts as a Senior UX Designer. It looks at your website copy and "Reverse Engineers" who the customer is and what they are trying to do. It then tells you how to fix your navigation and layout to convert them better.
Tags: UX, Website, Design
True comprehension is achieved not by passive reading, but by teaching. Instead of the AI lecturing you, you lecture the AI. The AI will act as a beginner, pointing out your reliance on jargon, asking for analogies, and finding holes in your logic.
Tags: Feynman, Learning, Teaching
Perfect for professional upskilling, corporate training, or mastering new software. Instead of giving you the "TL;DR", this prompt forces critical thinking by absolutely refusing to give direct answers. It coaches you to the finish line.
Tags: Socratic, Education, Questioning
Passive reading leads to rapid forgetting. This prompt extracts the core mechanisms of a document and tests you on them to enforce long-term retention. The output is formatted specifically so you can copy-paste it into flashcard apps like Anki or Quizlet.
Tags: Flashcards, Quiz, Retention
This is a heavy-duty QA tool. It forces a "3-Pass Protocol" where the AI acts as a destructive red teamer before allowing itself to write the final version. It is specifically tuned for NZ cultural context (macrons) and spelling.
Tags: QA, Red Team, NZ English
This utilizes "Chain of Verification" logic. By forcing the AI to critique its first attempt before you see the final result, you get a much higher quality output. It's like asking the AI to "think before it speaks."
Tags: Self-Correction, Chain of Thought, QA
Creates a "Clone" of your writing voice. It specifically looks for your "Negative Constraints" (words you never use) which is the secret to sounding human.
Tags: Style, Voice, Clone
Checks if your writing is boring. It ensures abstract concepts (like "Synergy") are anchored in concrete realities (like "Saving 5 hours a week").
Tags: Story, Narrative, Audit
Uses the AI to fix your bad prompts. If you can't get the result you want, ask the AI to rewrite your request using the principles of this library.
Tags: Prompt, Optimisation, Meta
Capture lightning in a bottle. If you have a long, messy chat that eventually gave you a perfect answer, this prompt reads the history and creates a single "Master Prompt" so you can get that result instantly next time.
Tags: Prompt, Reverse Engineer, Learning
Refine text specifically for syntax, grammar, readability, and stylistic consistency without altering the overarching message.
Tags: Persona, Editing, Copy
Evaluate the overarching macro-structure, pacing, conceptual strength, and audience alignment of a manuscript or proposal.
Tags: Persona, Structure, Editing
Rigorously stress-test systems, plans, architectures, or codes to proactively identify vulnerabilities, edge cases, and systemic leaks before they are deployed.
Tags: Persona, Red Team, Security
Challenge the user's assumptions, provoke deeper critical thinking, and refine raw ideas through an ongoing oppositional dialogue.
Tags: Persona, Debate, Critical Thinking
Analyse text, arguments, or transcripts specifically to identify cognitive biases, structural flaws, and invalid reasoning.
Tags: Persona, Logic, Analysis
Verify claims, statistics, and historical data against established knowledge bases to ensure absolute accuracy and credibility.
Tags: Persona, Fact-Check, Verification
A clear-eyed advisor for navigating workplace and personal politics with integrity. Maps stakeholders and underlying interests, names the real tension (and your own part in it), and gives you 2–3 honest paths plus a conversation plan. Refuses manipulation and redirects harassment, discrimination, or legal matters to HR/EAP.
Tags: Persona, Stakeholders, Difficult Conversations
Applies a structured Te Tiriti impact lens to a proposal — Articles 1–3, principles overlay, and an engagement map. Designed to surface gaps without inventing iwi positions or hollow performative engagement.
Tags: Te Tiriti, Public Sector, Equity
Triages an inbound OIA / LGOIMA request, classifies scope and transfer, calculates deadlines, proposes withholding grounds with public interest balancing, and drafts a defensible reply plus internal file note.
Tags: OIA, Public Sector, Ministerial
Stress-tests a draft Cabinet paper against CabGuide structure, consultation requirements, Tiriti analysis, RIS, financials, and proactive release. Surfaces what will get flagged at agency check before it gets there.
Tags: Cabinet, Public Sector, Policy
Treasury-style 5-Case Model audit of a Better Business Case. Scores each case, distinguishes asserted from evidenced claims, and lists the specific issues that will fail at Investment Panel.
Tags: Treasury, Business Case, Investment
Screening PIA against the Privacy Act 2020 and IPPs 1–13. Determines whether a full PIA is needed, heatmaps IPP risks, and proposes design-time mitigations including Māori data sovereignty considerations.
Tags: Privacy, Compliance, Public Sector
Checks a proposed AI/agentic workflow against the Algorithm Charter for Aotearoa NZ and Public Service AI expectations. Risk-tiers the use case, scorecards transparency and oversight, and drafts a plain-English public explanation.
Tags: AI Governance, Public Sector, Algorithm Charter
Reviews a draft procurement plan or RFx against the Government Procurement Rules, the Charter, and Progressive Procurement targets. Flags compliance gaps, distinguishes genuine broader outcomes from tokenistic ones, and proposes RFx-level edits.
Tags: Procurement, GETS, Public Sector