/lynqu-deal-desk

name: lynqu-deal-desk
description: Price a Lynqu deal and draft the quote. Browse the price book, put priced lines on a deal, then produce a draft quote for a human to send. Requires the Lynqu MCP server connected.

Lynqu Deal Desk

You price a deal from the org's own price book, put the lines on the opportunity, and produce a **draft** quote. A human sends it. That boundary is deliberate and absolute: sending a quote moves money and creates a commitment, so it stays with the person whose name is on it.

Two rules of the object model that trip everyone up, stated once:

1. **A deal's value is the sum of its line items.** You do not set it by hand. Add, remove or reprice the lines and the value follows. A hand-set value that contradicts the lines is the number that shows up wrong in the forecast. 2. **A price-book item can only go on a deal in the same currency.** A EUR deal takes EUR catalog items. Mismatches are refused, and correctly so: silently converting someone's price list is how a margin disappears.

Invocation

/lynqu proposal <deal | lead | company>
/lynqu quote <deal>

Step 1: Find the deal

• **`list-opportunities`** / **`get-opportunity`**: the deals on the account.

One lead can carry several: a renewal and an expansion are separate deals with separate values, and merging them in your head produces a quote nobody can approve

• **`get-lead`**: stage, score, the notes that say what was actually agreed
• **`list-quotes`** / **`get-quote`**: **read this before drafting anything.** A

quote already in flight means you're revising, not creating, and a second independent quote for the same scope is a real problem in front of a customer

• **`get-company`**: who's actually buying, for the quote header

If no opportunity exists yet, **`create-opportunity`** on the lead, but only when scope is genuinely understood. A deal created to have something to quote against becomes a forecast entry nobody believes.

Two clean-ups belong here, before any pricing, and each needs its own yes:

• **A secondary deal that isn't real** (a renewal that was never going to

happen, an upsell opened twice): **`delete-opportunity`**, naming the deal. A lead's primary deal cannot be deleted; it closes with a stage move on the lead

• **The account split across two company records** (the roll-up on

`get-company` is missing deals you know exist): **`merge-companies`** with no arguments lists the duplicate pairs with each side's lead, contact and deal counts. Merging moves everything to the survivor and cannot be undone here, so show both sides and confirm that one pair. Needs `companies.manage` (managers by default)

Step 2: Confirm the scope before you price it

The most expensive failure in this skill is a clean quote for the wrong thing.

Read the notes and contact points for what was agreed: how many seats, which modules, what term, when it starts. Where the record is ambiguous, **ask**. One question now beats a revision cycle and a credibility hit later.

Explicitly confirm:

• **Quantity** and what it's measured in (seats? locations? events?)
• **Term**: monthly, annual, multi-year
• **Start date**: especially if it's tied to an event or a renewal
• **Currency**: the deal's currency, which the catalog items must match
• **Anything already promised** in a conversation. A discount mentioned verbally

and missing from the quote is a fight in week two

Step 3: Build the lines from the price book

• **`list-catalog-items`**: the org's price book. Filter to the deal's currency
• **`manage-deal-line-item`**: add each line: item, quantity, and any agreed

discount. Name and price are snapshotted onto the line at add time, so a later price-book change doesn't silently rewrite a sent quote

• Re-read **`get-opportunity`** afterwards and confirm the rolled-up value is what

you expect. If it isn't, a line is wrong. Don't "fix" it by overriding the value

**Discounts.** Apply only what's been agreed and record why on the line or in a note. An unexplained discount is a precedent: the next deal in that segment gets asked for the same one, and nobody remembers the reason.

**Missing from the price book?** Say so. Don't invent a price. Either the item needs adding by someone with the authority (`manage-catalog-item`, admin territory in most orgs) or the deal needs custom pricing that a human has to own.

Step 4: Draft the quote

• **`manage-quote`**: create or revise the draft: line items, validity period,

terms, the customer's details

• Versioning is handled server-side. A revision supersedes its predecessor rather

than overwriting it, so the history of what was offered survives. Don't try to edit a sent quote in place

• Once a quote is **sent**, its contents are frozen. Changes mean a new version.

That's a feature; treat it as one

• The quote renders in the lead's locale where set, so the customer reads it in

their own language. Check the currency and locale look right before handing it over

**Sending, accepting, declining and revising a sent quote are deliberately not available here.** They email customers and move deal value. Point the user at the Lynqu app: this is human-in-the-loop by design, not a missing feature, and saying that clearly is better than apologising for it.

Step 5: Set up what happens next

• **`add-lead-contact-point`**: a dated follow-up on the quote. Quotes go stale

in silence; a task two days after sending is what converts one

• **`add-lead-note`**: the pricing rationale: what was agreed, what was

discounted and why, what was deliberately excluded

• **`update-lead-stage`** if the deal genuinely moved to a proposal stage
• **Forecast category**: set it honestly. `get-forecast` is only as good as the

categories under it, and an optimistic commit is a promise someone else has to keep

Output format

# Quote Draft · {Company}

**Deal:** {name} · {currency} · {term} · starts {date}
**Statu