Version 1.0 | Personal Use Only
Paste directly into Claude, ChatGPT, or any LLM. Replace[brackets]with your specifics.
Category: Productivity & Focus
I have [X hours] today and the following tasks: [list your tasks]. My energy peaks in the [morning/afternoon/evening]. Design a time-blocked schedule that protects 90-minute deep work sessions, batches shallow work (email, admin, calls) together, and includes 15-minute buffers between major blocks for unexpected interruptions. Output as a clean hourly calendar with task names and rationale for each block.
Category: Productivity & Focus
I’m stuck on a decision: [describe the decision]. I keep going back and forth. Help me think through this using a structured framework. Identify the core tradeoff, list the top 3 risks of each path, note what information I’m missing that would make this easier, and give me a clear recommendation with your reasoning. Don’t hedge — just help me decide.
Category: Productivity & Focus
It’s the end of my week. Here’s what happened: [brief summary of your week — wins, losses, tasks completed, things that slipped]. Run a structured weekly review: What went well and why? What underperformed and why? What should I stop, start, or continue? What are the top 3 priorities for next week? Output as a concise review doc I can save and reference.
Category: Code & Dev
Review the following code as a senior engineer who cares equally about readability, security, and performance. For each issue you find: state the problem clearly, explain why it matters in production, and provide the corrected code snippet. Don’t just flag issues — fix them. If the code is solid, say so and explain why it’s good. Here’s the code:
[paste your code here]
Category: Code & Dev
I’m debugging an issue and I’m stuck. Here’s the context: Language/framework: [X]. What the code is supposed to do: [describe expected behavior]. What it’s actually doing: [describe the bug, include error messages]. What I’ve already tried: [list your attempts]. Walk me through your reasoning step by step. Don’t just give me the answer — help me understand what’s wrong and why, so I can recognize this pattern next time.
Category: Code & Dev
I’m designing a system to [describe what you’re building]. Current constraints: [tech stack, team size, scale requirements, timeline]. Walk me through two or three architectural approaches that would work, compare their tradeoffs honestly (not just the happy path), and recommend which one fits my constraints best. Flag any common mistakes you’d expect a team at this stage to make with each approach.
Category: Business & Marketing
Write a cold outreach message for the following situation: I’m reaching out to [target: job title/company type]. My offer: [what you do or sell]. The problem I solve: [specific pain point]. My credibility: [social proof, past results, or relevant background]. Write 3 versions — one email, one LinkedIn DM, one Twitter/X DM. Each should be under 100 words, lead with their problem (not my pitch), and end with a low-friction CTA. No buzzwords. No “I hope this finds you well.”
Category: Business & Marketing
My product/service: [describe it]. My current positioning: [your current tagline or pitch]. My target customer: [who they are, their job, their pain]. My top competitors: [list 2-3]. Help me sharpen my positioning. Identify where my current pitch is generic or weak, suggest 3 positioning angles that differentiate me, and write a revised one-liner for each angle. Then pick the strongest one and explain why.
Category: Business & Marketing
I’m launching [product/service/feature] in [X days/weeks]. My audience: [size and where they live — email list, Twitter, etc.]. My goal: [revenue target, sign-ups, etc.]. Build me a launch sequence with specific actions for each day of the final 7 days before launch. Include what content to publish, what to tease vs. reveal, when to send emails, and how to create urgency without being spammy. Give me a day-by-day plan I can execute.
Category: Writing & Creativity
I’m writing a [blog post / essay / article] about [topic]. My target reader: [describe them]. My thesis or main argument: [what you want to prove or explore]. My tone: [conversational / authoritative / provocative / etc.]. Write a detailed outline with H2 and H3 headers, a suggested opening hook, and a paragraph of guidance for each major section. Don’t write the full piece yet — just give me a scaffold strong enough to build on.
Category: Writing & Creativity
Here are 3 examples of my writing: [paste 3 short samples]. Analyze my voice: identify my sentence rhythm, vocabulary level, use of metaphor, level of formality, and any signature patterns. Then rewrite this draft in my voice: [paste the draft you want matched]. Keep the substance identical — just make it sound like me, not like a generic AI.
Category: Writing & Creativity
I’m brainstorming [content / product names / story ideas / campaign concepts] for [context]. Give me 20 raw ideas — don’t filter for quality yet, just generate. After the list, pick your top 3, explain what makes them strong, and develop each one into a 2-sentence concept. I want range: some safe, some weird, some high-risk-high-reward.
Category: AI Agents & Automation
I’m building an AI agent to [describe the agent’s job in one sentence]. It will have access to: [list tools or data sources]. Its output should be: [format — JSON, markdown, plain text, etc.]. Write a complete, deployable system prompt that defines the agent’s role, its constraints and guardrails, how it should handle ambiguous input, what it should refuse, and what its output format should look like. Make it specific and tight — not a generic “helpful assistant” boilerplate.
Category: AI Agents & Automation
Here’s a complex task I want to automate: [describe the task end-to-end]. Break it into a multi-step workflow that could be executed by an AI agent or a series of AI agents. For each step: name the step, describe the input it receives, describe the output it produces, identify what tool or capability it needs, and flag any steps that require human-in-the-loop review. Output as a numbered workflow diagram in plain text.
Category: AI Agents & Automation
Here’s a prompt I’ve written for an AI agent: [paste your prompt]. I need you to stress test it. Generate 10 edge case inputs that might break it, confuse it, or produce unexpected outputs. For each edge case, explain why it’s a problem and suggest a revision to the original prompt that would handle it gracefully. Then give me a revised, hardened version of the full prompt that incorporates your fixes.
Category: AI Agents & Automation
Here’s a description of how I currently do [task or process]: [describe your current manual workflow step by step]. Analyze this workflow and identify: the 3 highest-leverage steps to automate first, what tool or AI capability would handle each, the estimated time saved per week, and any risks or edge cases I should handle before automating. Rank your recommendations by ROI — biggest time savings with lowest implementation complexity first.
Category: AI Agents & Automation
I need to design the persona and behavior profile for an AI agent that will interact with [end users — customers, team members, etc.] in the context of [describe the product or platform]. Define: the agent’s name and role, its tone and communication style, 5 example responses to common user messages, what it should never say or do, and how it should handle escalations or situations outside its scope. Make it feel like a real, consistent personality — not a corporate script.
Category: AI Agents & Automation
I want to complete [feature/task] with AI while using the right model tier for each phase. Design a 3-step workflow: (1) requirements gathering with a cheap fast model like Flash, Mini, or Haiku-level to extract goals, missing context, assumptions, edge cases, and acceptance criteria; (2) planning with a Sonnet-level medium premium model to produce the technical plan, dependency order, risks, and verification checklist; (3) execution with a Sonnet-level coding model by default. Only recommend a higher-reasoning model if Sonnet-level execution fails twice for the same reason or the task is genuinely too complex/long-running for Sonnet-level reasoning. Include a mandatory final code review step, and make it clear that code review matters more than planning because review catches the mistakes that actually ship.
End of starter prompt library. Total: 18 prompts across 5 categories.
License reminder: These prompts are for your personal use only. No resale, no redistribution, no sharing with teams or clients without purchasing additional licenses. See LICENSE.md.