CounselKit — 24 AI Prompts for Legal Professionals. Premium Prompt Pack. $49 one-time.

Legal prompts for how you actually work.

Get CounselKit — $49
Instant PDF download · Works with ChatGPT, Claude, and any AI tool

The prompt problem.

Most lawyers use AI every day now. The results are usually mediocre. Not because the tools are bad — ChatGPT and Claude are remarkably capable. The problem is that they start every conversation knowing nothing about you, your work, or the standards your profession demands. So they guess. They cite cases that don't exist — over 1,100 court cases worldwide now involve fabricated AI citations, and 128 U.S. lawyers have been sanctioned. They draft provisions that sound right but miss what matters.

As one attorney put it: "Every time I demo one of these legal-specific tools, they suck." Another described "countless hours trying to improve the legal research capabilities of ChatGPT through clever prompting" — and failing.

The difference between a useless AI response and one that actually saves you time is almost always the prompt. Not a better subscription. Not a fancier tool. Just a better question, asked with the right structure.

CounselKit is 24 prompts built specifically for legal workflows. Each one tells the AI who it is, what to do, how to format the result, and where to flag uncertainty instead of guessing. Copy one into ChatGPT or Claude, fill in the brackets, and get output that's actually useful.


What's Inside

01

Foundation

The CREATE Framework · Prompts 01–05

Five prompts that teach the AI how to work with you. Set its role, load context, specify output format, iterate on results, and constrain scope. Use these before every major task — they make everything else work better.

02

Document Review

Prompts 06–12

Compare documents side by side. Summarize key provisions. Identify what's missing. Tier issues by risk. Translate dense clauses into plain English. Explain redlines. Seven prompts for the work that fills most of your day.

03

Research & Analysis

Prompts 13–18

Summarize sources without hallucination. Map competing arguments. Identify gaps in your analysis. Compare jurisdictions. Check citations. Synthesize meeting prep from scattered materials. Six prompts for thinking clearly with AI assistance.

04

Drafting

Prompts 19–24

Generate structured first drafts. Build a clause library. Audit defined terms for consistency. Codify your writing style so the AI matches it. Run a QA protocol before anything goes out the door. Revise to client-ready. Six prompts for producing work product, not rough drafts.

Plus

Customization Section

A prompt builder to create your own prompts using the CounselKit framework, plus drop-in AI profile setup files for ChatGPT Custom Instructions and Claude Project Instructions. Set them once — every conversation after that starts with your professional context already loaded.

See the prompts before you buy.

These are two of 24 prompts in the pack. Every prompt includes the full template, a description of when to use it, and a pro tip from real legal workflows.

09
Risk Tiering

Categorize identified issues by severity — High, Medium, Low — with clear reasoning so you know where to focus your effort.

I've identified several issues in a document (or you can identify them first using the Issue Spotter prompt). Now I need you to tier them by risk severity. **Here are the issues:** [PASTE YOUR LIST OF ISSUES HERE, or] [ATTACH THE FULL DOCUMENT OR PASTE ITS TEXT HERE — identify and tier issues from scratch] --- **Your task:** For each issue identified, assign a risk tier and explain why. **Use this tiering framework:** **HIGH RISK:** Exposure to significant liability, financial loss, operational disruption, or legal violation; immediate attention required. **MEDIUM RISK:** Potential exposure that's non-trivial but manageable; should be addressed before execution but not blocking. **LOW RISK:** Clarification or optimization opportunity; nice to address but not a blocker. --- **Format your response as:** **HIGH RISK** - [Issue title] — [1-2 sentence explanation of why this is High] - Specific exposure: [What goes wrong if we proceed as written?] - How to fix: [Minimal suggestion for remediation] **MEDIUM RISK** - [Issue title] — [1-2 sentence explanation] - Specific exposure: [What goes wrong?] - How to fix: [Suggestion] **LOW RISK** - [Issue title] — [1-2 sentence explanation] --- **Critical instruction:** Do NOT assign a risk tier without explaining specifically what makes it High/Medium/Low. A bare assertion ("This is high risk") is worthless. You must articulate the actual downside. If you're uncertain about the risk level, note that: "This appears High Risk, but I'm flagging uncertainty about [specific gap]." --- **My context (if relevant):** - Type of agreement: [e.g., "service contract"] - My role: [e.g., "buyer," "service provider"] - Key concerns: [e.g., "liability exposure," "confidentiality"]

Pro tip: Risk tiering is discipline. It forces you to be specific about downside risk instead of vague worry. Use this rigorously.

12
Redline Explainer

Given a redlined document, explain what changed, why it matters, and what's worth pushing back on.

I have a redlined document. I need you to explain the redlines: what changed, what it means, and what (if anything) I should push back on. **The redlined document (show changes with + for additions, - for deletions, or paste with track changes visible):** [ATTACH THE REDLINED DOCUMENT OR PASTE THE TEXT WITH CHANGES MARKED] --- **Your task:** For each substantive redline: 1. Explain what changed (in plain English) 2. Why it might matter 3. Whether it's a real issue or minor wordsmithing 4. What to do about it (push back, accept, propose alternative) **Format as:** **Redline 1: [Section/clause affected]** - What changed: [Original text] → [New text] - What it means: [In plain English, what's the practical effect?] - Why they made this change: [Their likely motivation] - Does it matter? [HIGH PRIORITY / MEDIUM / LOW / NONE] - My recommendation: [Accept / Push back / Propose alternative — and why] [Repeat for each redline] --- **Summary:** - Total redlines: [Count] - Must fight: [Which are deal-breakers?] - Can live with: [Which are acceptable?] - Room to negotiate: [Which can you trade for something else?] --- **Critical instruction:** - Flag when the significance of a change is ambiguous. Don't guess. - If a redline is tied to other provisions or creates contradictions, note that. - Be strategic: note which redlines are worth your negotiating capital and which aren't. - If a redline is unclear, say so instead of guessing.

Pro tip: Most redlines are negotiable. Knowing which hills to die on is more important than fighting every change.


Three steps. No setup.

1

Download the PDF

58 pages, organized by workflow. Save it, print it, keep it open on a second monitor. It's yours.

2

Copy a prompt into your AI tool

ChatGPT, Claude, Gemini — whatever you use. Each prompt is self-contained and tool-agnostic.

3

Fill in the brackets and go

Every prompt has clear [PLACEHOLDER] fields. Paste your document, describe your situation, get structured output you can actually use.


Not another prompt list.

Built for legal work, not adapted from marketing templates

Every prompt in CounselKit was written for the way lawyers actually work — document review, issue spotting, risk analysis, client communication. These aren't generic "write me an email" prompts with the word "legal" bolted on.

Anti-hallucination guardrails are built into every prompt

Stanford researchers found general-purpose AI hallucinates 58–82% of the time on legal research queries. Even RAG-enhanced tools like Lexis+ AI still produce errors 17% of the time. CounselKit prompts instruct the AI to flag uncertainty, refuse to fabricate citations, and distinguish between what the law clearly says and what requires verification.

A workflow system, not a wall of text

The prompts are organized into four modules that mirror actual legal workflows: set up the AI, review documents, research and analyze, then draft. Module 1's foundation prompts work as a base layer for everything that follows. This structure means you learn a system, not just a list of tricks.

Less than a single billable hour

Harvey AI costs ~$1,200/seat/month. CoCounsel runs $225–$500/month. CounselKit is $49 once. It works with the ChatGPT or Claude subscription you already have.


Built for practitioners.

This is for attorneys, in-house counsel, paralegals, and legal ops professionals who use AI tools in their daily work and want dramatically better results. You don't need to be technical. You don't need to understand "prompt engineering." You need prompts that respect how legal work actually functions — the precision it demands, the uncertainty it requires you to flag, and the professional standards it has to meet.

If you've ever pasted a contract into ChatGPT and gotten back something that sounded confident but was mostly wrong, CounselKit is for you.


$49

One-time purchase. No subscription.

24 structured prompts across 4 workflow modules · Custom prompt builder · AI profile setup files for ChatGPT and Claude · Free updates

Get CounselKit — $49

Not useful in your first week? Email me and I'll refund you. No questions asked.