Legal & Compliance

Verified against ChatGPT · 2026-08-08

Flag the clauses in a draft contract that carry outsized risk before it goes back to the other side

Scans a contract draft for clauses that create disproportionate risk relative to the deal size, ranks them by exposure, and produces redline talking points — framed as a starting point for negotiation, not a legal opinion.

ChatGPT (GPT-5.1)4 fillable variables
Scope for this category: Drafting, summarizing and organizing support only — every prompt states plainly that output is not legal advice and needs review by a qualified lawyer before being relied on or sent externally.

The prompt

Ready to copy — highlighted parts are example details you can swap.

Review the contract draft below and identify clauses that create risk disproportionate to the size or nature of this deal. This is risk-flagging to prepare for negotiation, not a legal opinion on enforceability.

CONTRACT DRAFT
Draft master services agreement from a vendor, sections on liability, IP, and payment terms.

DEAL SIZE AND NATURE
$40,000 annual contract for a marketing automation tool, one-year term.

OUR SIDE OF THE DEAL
We're the customer; main concern is not being liable for the vendor's data breach.

KNOWN DEALBREAKERS
We will not accept uncapped liability under any circumstances.

PHASE 1 — SCAN
Go through the draft clause by clause and identify every provision that shifts risk, cost, or obligation onto our side disproportionately to the deal size stated above — uncapped liability, broad indemnification, one-sided termination rights, unusually long payment terms, IP assignment beyond what the deal requires, or auto-renewal without an out. Do not flag a clause just because it's unfavorable in the abstract; flag it because it's unfavorable relative to what this specific deal is worth to us.

PHASE 2 — RANK
Rank the flagged clauses by financial or operational exposure, highest first. For each one, state in one sentence what could actually happen if it were invoked against us, not just what the clause says.

PHASE 3 — TALKING POINTS
For each flagged clause, draft one negotiation talking point — a specific ask (cap the liability at X, mutual termination rights, narrow the indemnification scope) rather than a vague objection like "this seems risky." If a known dealbreaker is present, mark it separately as non-negotiable rather than a talking point.

WHAT NOT TO DO
Do not state whether any clause is enforceable or unenforceable under any jurisdiction's law — that determination depends on facts and law you have not verified and should not assert. Do not invent a specific statute, case, or regulation to justify a flag; if a legal basis matters, say that it should be confirmed by counsel rather than naming one.

OUTPUT FORMAT
1. Ranked table: Clause | What it does | Exposure if invoked | Talking point.
2. Separate short list of anything matching a known dealbreaker.
3. A closing line stating this is a draft risk-flagging pass to prepare for negotiation, not a legal opinion, and that the marked clauses — especially anything involving indemnification, liability caps, or IP — should be reviewed by a qualified lawyer before the contract is finalized or signed.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

A generic "find the risky clauses" instruction produces a flat list because the model has no anchor for what counts as disproportionate, so it either flags everything that sounds legally serious or nothing at all; requiring the deal size and the user's specific side as inputs gives GPT-5.1 a concrete basis for relative judgment — a broad indemnification clause is a minor flag on a $2,000 deal and a major one on a $2 million deal, and the model can only make that distinction if the comparison point is stated rather than implied. Separating the scan phase from the ranking phase keeps the model from prematurely deciding a clause isn't worth mentioning while it's still cataloguing — a common failure mode where an LLM's first pass at "is this important" quietly drops borderline items before a ranking step ever gets to weigh them against each other. Forcing each flagged item into a specific negotiation ask rather than a general objection matters because "this seems risky" gives the person going back to the other side nothing to actually propose, while "cap liability at 12 months' fees" is something that can be pasted into a redline comment directly. The explicit ban on asserting enforceability or citing a specific law is the load-bearing safety constraint here: risk-flagging for negotiation prep is a business-judgment task the model can reasonably help with, but the moment it states a clause is unenforceable under some jurisdiction's law it has crossed into legal opinion territory without the facts, jurisdiction confirmation, or license to back it up.

What you get back

1. Indemnification (Sec. 8) — vendor requires us to indemnify them for any claim arising from our use of the product, uncapped. Exposure: could exceed contract value many times over in a serious incident. Talking point: propose a mutual indemnification cap at 2x annual fees. 2. Auto-renewal (Sec. 14) — renews for a full year unless cancelled 90 days out. Talking point: shorten notice window to 30 days. Dealbreaker match: none of the above conflicts with your stated non-negotiables. This is a draft risk-flagging pass, not a legal opinion — have a lawyer review the indemnification and liability sections before signing.

Verified against

ChatGPT GPT-5.1 · 2026-08-08

Changelog

  • 2026-08-08 Initial publish, verified against ChatGPT GPT-5.1.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS
All Legal & Compliance prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY