Write a GPT Store listing that gets picked over the ten other GPTs like it

You are writing the public GPT Store listing for a Custom GPT already built and configured — the name, description, category, and conversation starters a stranger browsing or searching the Store will actually see before opening the GPT itself.

WHAT THE GPT ACTUALLY DOES
Reads an uploaded lease or vendor contract PDF and explains each clause in plain English, flagging which clauses are unusually one-sided compared to standard templates — no live legal-database lookup, purely reasoning over the uploaded document.

WHO IT'S REALLY FOR
Small business owners signing their first commercial lease or vendor contract without a lawyer on retainer.

WHAT MAKES IT DIFFERENT FROM SIMILAR GPTS
Most "contract review" GPTs in the Store just summarize; this one specifically flags clauses that deviate from standard boilerplate, which is the actual thing a non-lawyer can't judge on their own.

CATEGORY IT SHOULD LIST UNDER
Considered "Productivity" but "Legal" is the more accurate fit even though Productivity likely has more browse traffic.

LISTING RULES
Write the name to describe the specific outcome the GPT produces, not a vague capability word — a name like "Contract Clause Explainer" tells a searcher exactly what they'll get, while something like "Legal Helper" reads as one of dozens of similarly generic names already crowding the same category and gives a browsing user no reason to open this one over the next result. Write the description's first sentence to state the specific differentiator, not a restatement of the category — a searcher scanning a results page reads only the first line of most listings before deciding whether to open one, so if the actual point of differentiation is buried in sentence three, it functions the same as not having a differentiator at all for anyone who doesn't click through. Do not claim a capability the GPT doesn't actually have configured — if it doesn't have a live-data Action wired up, don't write a description implying it checks real-time information, since a listing overselling a capability produces a one-star review the moment a user tests the exact thing the description promised and the GPT can't do it. Write four conversation starters that are the actual first messages a real target user would type, in their own likely phrasing, not generic prompts — each one should imply a different real use case so a browsing user can tell from the starters alone whether their specific situation is covered. Pick the single most accurate category rather than the highest-traffic one if they differ — a GPT that's genuinely a writing tool mis-categorized under Productivity to catch more browse traffic gets discovered by the wrong audience and reviewed against expectations it was never built to meet.

OUTPUT FORMAT
1. Name, with a one-line note on why it beats a more generic alternative considered and rejected.
2. Description, front-loaded with the differentiator in its first sentence.
3. Recommended category.
4. Four conversation starters in the target user's own likely phrasing.