Pull the lesson from a closed deal before it fades.
Reconstructs the timeline, separates what moved it from noise, names the near-death moment, and lands one repeatable lesson.
Inside: The timeline, cause vs noise, and one portable lesson.
Install it in one line, or paste it in.
~/.claude/skills/ and runs automatically when it is relevant.Connect your context. Set it to your motion.
it reads the stage history, close date, amount, and the activity trail automatically, so the timeline is real, not remembered.
pulls what was actually said at the turning points, so the debrief is grounded in the record, not the story you tell yourself.
on trials, shows whether usage matched the story the deal told.
This was built for a B2B SaaS org running a staged pipeline. Set these to your stack:
The point is one lesson you will actually reuse, not a form nobody reads. Set the capture target to where your team will actually see it.
| Set this | What it is | Default / Example |
|---|---|---|
| STAGE history | how you see the deal's path | Opportunity stage historyclose date |
| OUTCOME field | won, lost, or no-decision | StageNameClosed Won / Closed Lost |
| LOSS_REASON field | why it closed the way it did | Loss ReasonCompetitor field |
| ACTIVITY sources | where the touch trail lives | TaskEventmeeting or email tool |
| CAPTURE target | where the lesson is written | your notesa shared doca deal wiki |
| LESSON_SCOPE | how broadly a lesson should apply | this repthis segmentthe whole team |
Everything the skill does, in full.
Takes one closed deal, won or lost, and pulls the lesson out of it before it fades. It rebuilds the timeline from first touch to close, separates what actually moved the deal from what only felt important, names the moment it nearly died, and lands on a single repeatable lesson worth carrying to the next deal. Then it captures that lesson in a form the team can reuse, so the same mistake is not paid for twice.
- 1Reconstruct the timeline (what happened, in order)
Lay the deal out from first touch to close: the stages, the dates, the key conversations, the moments something changed. Use the record where you have it. A debrief built on memory alone remembers the story, not the deal.
- 2What actually moved it (separate cause from noise)
Name the two or three things that genuinely moved the deal toward its outcome. Strip out what felt important at the time but changed nothing. A win from a warm intro is a different lesson than a win from a great demo. Be honest about which it was.
- 3What nearly killed it (the near-death moment)
Every deal has a moment it almost went the other way. Name it. What caused it, what recovered it or failed to, and whether you saw it coming. This is usually where the real lesson lives.
- 4The one repeatable lesson (single, portable)
Land on exactly one lesson that would change how the next deal is run. Not five. One, stated so plainly that another rep could apply it without you in the room. If you cannot say it in a sentence, it is not the lesson yet.
- 5Capture it (so it is not relearned)
Write the lesson where the team will actually find it, tagged to the situation it applies to. A lesson that lives only in one rep's head gets paid for again by the next rep. Capture turns one deal's cost into everyone's edge.
- The timeline is built from the record where it is available, and clearly marked where it rests on memory.
- What moved the deal is separated from what only felt important, honestly, even when the honest answer is luck.
- The debrief lands on exactly one repeatable lesson, stated portably, not a list of five.
- Loss reasons name what actually happened, never a tidy category that hides it.
- Any date or amount carried from the deal is shown as given or sourced, never invented.
DEAL DEBRIEF · Acme Corp · CLOSED WON · $60K (illustrative) Timeline Day 0 Inbound from a webinar Day 12 Demo to VP Ops (the eventual champion) Day 34 Stalled in legal for 3 weeks Day 61 Signed after an exec-to-exec call What actually moved it 1. The champion sold it internally while we were out of the room. 2. The exec call broke the legal stall. The demo did not close it. What nearly killed it The legal stall. We had no second thread and no exec sponsor until day 55. We got lucky the champion held on. The one repeatable lesson Line up the exec sponsor before legal, not after the stall. Waiting cost us three weeks and nearly the deal. Captured to: team deal wiki, tagged "mid-market, legal stall"
Where an operator takes this next.
The read is step one. Here is where an operator takes it once the manual version proves out.
One deal's cost, captured once, becomes the whole team's edge.
Wire a scheduled Claude task to fire whenever an opportunity flips to Closed Won or Closed Lost in Salesforce.
Connect the deal wiki to your CRM so a rep entering a similar deal sees prior lessons tagged to that situation.
Connect Gong so the near-death moment gets the actual transcript excerpt instead of a paraphrase.
One skill is the on-ramp.
A single skill does one job. Chained into a playbook, or run as a full build, it becomes a system. Here is where this one plugs in.