The best Claude skills for sales don't write a follow-up email and call it selling. They score the deal against real evidence, pull the objection you'll actually get out of transcripts you already have, and prepare the specific conversation you're about to have. These 10 free Claude Code skills cover the part of the AE job that happens after someone agrees to talk: honest deal reads, pre-call briefs, objection prep, negotiation coaching, and renewal risk you catch four months out instead of in the renewal month.
Pipeline reviews run on optimism. A deal is "looking good" until the week it isn't, and the reason it slipped was almost always visible on a call three weeks earlier and never written down. The AE didn't miss it - they heard it, filed it as a probably-fine, and moved on to the next dial.
The best Claude skills for sales make that read explicit. They score the deal against stated criteria with an estimated-versus-confirmed split, pull the real objections out of transcripts instead of your memory of them, and prepare the specific conversation you're about to have rather than a generic playbook. Ten skills, MIT-licensed, ordered so you know which one earns its install first.
For context on why the honest-read job matters, Gong Labs' research on win rates has shown that reps who confirm next steps in writing at the end of a call close at materially higher rates than reps who don't - the confirmation itself is the tell, not the tactic. And Salesforce's State of Sales report puts reps at only 28 percent of their week actually selling; the rest is admin, research, and prep. That's the gap this pack closes. For the broader AE workflow this pack is built into, see gtm skills for account executives, and if you're comparing tools alongside these skills, see best sales engagement tools.
What is a Claude skill
A Claude skill is a versioned SKILL.md file with a stated method: the inputs it asks for, the framework it applies, and the shape of the output it returns. You install a pack once with npx skills add sidchaudhary/gtm-skills/skills/account-executive, restart Claude Code, and every skill in the folder is available by name. Two AEs running the same skill get the same output shape - a ten-dimension deal score with confirmed-versus-estimated flags, a five-objection playbook, a MEDDIC completeness table - so the work stops drifting between reps and between weeks.
Why AEs need Claude skills now
Two things are true at once for most AE teams. First, the forecast is mostly stories - "they're the champion, they're leaning in, it'll close this quarter" - and the ones that slip are the ones nobody scrutinized. Second, the pre-call research and post-call coaching that would catch it takes 30 to 45 minutes per deal, which reps don't have across a full pipeline. So the read doesn't happen, and the same deal slips a second quarter for a reason that was in a transcript from six weeks ago.
Claude skills fix this at the task level. The rep doesn't do more work; the mechanical parts of the work run through a skill with a fixed output shape. Winning by Design's research on rep productivity puts pre-call prep and post-call notes together at over a third of a rep's total selling time - that's the block a skill actually shortens.
The skills that move the needle for AEs answer three questions faster:
- Is this deal actually where I think it is, or am I forecasting on stories?
- What objection will I get on this call, and what am I saying back to it?
- Which renewal has runway to save, and which one is already a status report?
Install the pack in about 30 seconds
One command pulls all 10 Account Executive skills into your project's .claude/skills/ folder. Restart Claude Code afterward - skills load at startup, so a running session won't see them.
npx skills add sidchaudhary/gtm-skills/skills/account-executive
# restart Claude Code, then confirm:
# "What sales skills do you have?"You can install a single skill instead by appending its folder name, e.g. .../skills/account-executive/the-deal-gauge. The pack install is the better default because these skills reference each other's inputs - the-win-loss-analyzer feeds better objections into the-objection-playbook, and the-transcript-miner (in the GTM Engineer pack) feeds cleaner inputs into the-deal-gauge.
10 best Claude skills for sales
1. the-deal-gauge
Returns a health score and an intent score across ten weighted dimensions, plus a 7/14/30-day trend and a quadrant placement. The part that matters most: it flags which dimensions are confirmed versus estimated, so a 78 built on confirmed dimensions and a 78 built mostly on estimates are visibly different deals. It also outputs a MEDDIC (0-6) and BANT (0-4) completeness table and three to five prioritized next steps.
When to use: On the deal you'd describe as "probably fine." Not the obvious win, not the obvious loss - the one you haven't scrutinized because it isn't screaming at you.
How to run:
You: "Run the-deal-gauge on this deal."
[paste: notes, last call transcript, stage, close date]
Claude: -> health score + intent score across 10 dimensions
-> 7/14/30-day trend, quadrant placement
-> which dimensions are confirmed vs estimated
-> MEDDIC (0-6) and BANT (0-4) completeness tables
-> 3-5 prioritized next stepsBest used with: A real transcript or detailed notes. Stage-and-amount alone produces a score that reads like the CRM, which is the version you already didn't trust.
Where this skill hits its ceiling: One deal at a time. If you want the same read across the whole book, that's the-pipeline-scanner.
With Intempt: Deal scores run continuously off live product usage and CRM signal instead of being re-derived per pipeline review.
2. the-pipeline-scanner
Reads the whole pipeline and surfaces the deals whose stage doesn't match their evidence - stuck deals, risk signals, forecast by stage, coverage ratio. Its job is the one nobody wants to do before the QBR: name the deals that are lying to you.
When to use: The week before a forecast call, or any time your commit list has grown faster than the underlying activity would support.
How to run:
You: "Run the-pipeline-scanner on my current pipeline."
[paste: stage, amount, close date, last activity date per deal]
Claude: -> stuck deals: stage vs days-since-activity
-> risk signals per deal, ranked
-> forecast by stage with coverage ratio
-> the 3-5 deals to actually work this weekBest used with: Real activity dates. A CRM where every deal was "touched" last week because someone did a bulk update produces a scan that reads clean and isn't.
Where this skill hits its ceiling: It flags the lying deals; it doesn't work them. The-deal-gauge is the follow-up on the specific ones it names.
With Intempt: Pipeline health is a live view built off product-usage and engagement signal, not a Friday-afternoon export.
3. the-90-second-brief
Researches an account fresh from its name or domain - homepage, LinkedIn, press, job postings - and returns what they do, what changed recently, and the strongest angle to open with. Ninety seconds of read, not thirty minutes of tab-hopping. Meant to be the last thing you look at before the call.
When to use: Right before a first call with an account you haven't researched yet, or a re-engagement call after a quarter of silence.
How to run:
You: "Run the-90-second-brief on acme.com. First call in 20 min."
Claude: -> what they do, in one paragraph
-> what changed in the last 90 days (hiring, funding, press)
-> who's likely on the call and what they own
-> the strongest opening angle for this call, specificallyBest used with: A domain and a stated call purpose. "First discovery" and "expansion pitch to an existing customer" produce very different briefs.
Where this skill hits its ceiling: Public signal only. It doesn't know what happened on the last call - that's the-call-coach's prep side.
With Intempt: Account briefs pull first-party engagement data alongside public signal, so the "what changed" line includes what changed inside your product too.
4. the-call-coach
Does both sides of the call. Pre-meeting: the prep sheet with the questions you should ask given what you know. Post-meeting: a review that names what to change - talk ratio, BANT/MEDDIC scoring on the transcript, objections raised, and drills for the specific moment where the call went sideways. Specific, not "build more rapport."
When to use: Before every call you actually care about, and after every call that either surprised you or slipped. Not after every call - that's overhead.
How to run:
before: "Run the-call-coach prep for this account."
-> prep sheet with the questions you should ask
-> the objections most likely to come up
after: "Run the-call-coach review on this transcript."
-> talk ratio (yours vs theirs, by minute)
-> BANT/MEDDIC coverage scored 0-4 / 0-6
-> objections raised, and how you handled each
-> 2-3 specific drills for next timeBest used with: A real transcript for the review side. Notes work; transcripts work better because they catch the objection phrased the way the buyer phrased it.
Where this skill hits its ceiling: One call at a time. Pattern-finding across dozens of calls belongs to the-win-loss-analyzer.
With Intempt: Coaching runs continuously off recorded calls, and the drills feed back into rep-specific enablement instead of living in a doc nobody re-opens.
5. the-account-blueprint
Maps the buying committee for a named account and produces a phased 90-day action plan to get from your champion to the rest of the room. Champion, economic buyer, technical evaluator, blocker - who's identified, who's estimated, who's still a gap - plus the specific moves to close each gap.
When to use: When a named account is stuck at a single champion and the deal has stalled because nobody else in the room has entered the conversation.
How to run:
You: "Run the-account-blueprint on this stuck deal."
[paste: known contacts, last call notes, org info]
Claude: -> committee map (identified / estimated / missing)
-> who owns what, and who blocks what
-> phased 90-day plan: which stakeholder in which week
-> the specific ask that gets you to each new personBest used with: A deal where you already have a champion. Without a champion, the blueprint is a research doc, not a plan.
Where this skill hits its ceiling: It plans the moves; it doesn't run them. And a champion who won't multi-thread you can't be blueprinted around - that's a different conversation.
With Intempt: Committee mapping pulls product-usage signal per user, so who's actually engaged in the trial is visible next to who's on the org chart.
6. the-objection-playbook
Takes your product, ICP, and target persona and returns the five objections that persona raises most, each with an acknowledgment, a response, and a follow-up question. Paste the objections you actually hear on your calls and it tailors the playbook to those specifically instead of the generic five.
When to use: Before a call with a persona you're new to, or when the team keeps losing on the same objection and the response has been getting invented in real time.
How to run:
You: "Run the-objection-playbook.
Product: [yours]. Persona: VP Engineering at Series B SaaS.
Objections we actually hear: 'we're building this in-house',
'the timing isn't right', 'we already use [competitor]'."
Claude: -> 5 most likely objections for this persona
-> per objection: acknowledgment + response + follow-up Q
-> which objection is a stall vs a real blockerBest used with: Real objections you've heard, not the ones the manual lists. The generic version writes generic responses.
Where this skill hits its ceiling: It preps the answer, not the read. Whether you're getting a stall or a real objection is a judgment call on the day - the playbook makes both branches ready.
With Intempt: Objection patterns get mined from actual call transcripts instead of surveyed from reps, so the top-five list is the real top five.
7. the-switch-angle
Maps a prospect's current tool stack to the inefficiency it creates, and gives you the questions that surface that inefficiency on a call without you naming it first. Displacement work, done well. The point is a prospect who says the pain out loud - not a rep who lectures them about it.
When to use: On a competitive displacement where the prospect is running an incumbent tool and can't articulate why they'd switch.
How to run:
You: "Run the-switch-angle. Prospect uses [incumbent] for X.
Their stack around it: [tools]."
Claude: -> the inefficiency their current stack creates, named
-> 5 questions that get the prospect to name it
-> the moment on the call to introduce your alternative
-> the specific line that reframes cost of switchingBest used with: A named incumbent. "They use something" produces a generic angle; "they use [tool] alongside [tool]" produces a specific one.
Where this skill hits its ceiling: Prospects who genuinely love their current tool can't be switched by an angle - the skill will flag that instead of writing a fake one.
With Intempt: Stack signal gets pulled from the account brief automatically, so displacement angles show up as recommended plays without you feeding them in.
8. the-win-loss-analyzer
Ranks the real reasons deals close or die across a batch of closed-won and closed-lost deals, from evidence in transcripts and notes rather than rep memory. The team's story about why they lose is almost always partly wrong; the version built from transcripts contradicts it at least once, and that contradiction is the finding.
When to use: Once a quarter across the last 20 to 40 closed deals, before the retro starts running on vibes.
How to run:
You: "Run the-win-loss-analyzer on these 30 closed deals."
[paste or point at: transcripts, notes, stage history]
Claude: -> top reasons for won deals, ranked, with evidence
-> top reasons for lost deals, ranked, with evidence
-> which lost reason the team is under-counting
-> the 2-3 changes that would flip the biggest bucketBest used with: Real evidence per deal, even messy. Twenty deals with transcripts beat sixty deals with a one-line loss reason from the CRM.
Where this skill hits its ceiling: It names patterns; it doesn't fix them. The output is inputs into cold-outbound work (the SDR pack's the-cold-opener) and objection prep (the-objection-playbook), not a solved problem.
With Intempt: Win-loss patterns update continuously off closed deals instead of being a quarterly exercise, so the top-loss reason is current, not last quarter's.
9. the-renewal-tracker
Scores one existing account's renewal risk from contract value, days to renewal, a 90-day usage trend, and relationship signals, and names the single highest-impact action. Run per account while there is still runway - four months out it's a save, in the renewal month it's a status report.
When to use: On accounts renewing in 90 to 180 days, in a monthly pass. Not on renewals inside 30 days - by then the read isn't the useful move.
How to run:
You: "Run the-renewal-tracker on this account.
Contract: $X. Renewal: 120 days out.
Usage last 90 days: [trend]. Champion: [status]."
Claude: -> renewal risk score (0-100) with the drivers
-> usage trend read (declining, flat, growing)
-> relationship signal (champion changed jobs? gone quiet?)
-> the single highest-impact action, with a due dateBest used with: Real usage data. A tracker run without usage trend is a calendar reminder, not a risk score.
Where this skill hits its ceiling: One account at a time. If you want a recurring wrapper across the book, that's the-cadence-clock in the Lifecycle pack.
With Intempt: Renewal risk is a live score off product usage plus engagement, updated continuously - not a monthly rep exercise.
10. the-negotiation-coach
Preps the specific negotiation in front of you: anchor, concession ladder, walk-away point, and the calibrated questions that get the buyer to name their real constraint. The trade you decide before the call. A trade you improvise under pressure is a discount with extra steps.
When to use: Before any pricing, terms, or scope negotiation where the plan going in is currently "see what they say."
How to run:
You: "Run the-negotiation-coach on this deal.
List: $X. They asked for Y% off.
What we care about: multi-year, expansion path, logo."
Claude: -> anchor to open with, and why it's defensible
-> concession ladder: what you trade for what
-> the walk-away point, stated explicitly
-> 3-5 calibrated questions to surface their real constraintBest used with: A stated list price and a stated ask from the buyer. "They want a discount" is not enough input - the size and shape of the ask determines the whole ladder.
Where this skill hits its ceiling: It preps one negotiation. It doesn't set your pricing strategy - if the whole book is discounting 30 percent, that's a pricing problem, not a negotiation one.
With Intempt: Negotiation plays get informed by account-level signal (usage growth, expansion accounts, similar-deal history) instead of rep instinct alone.
Comparison: which Claude sales skill to use when
| Skill | Best moment to use | Time saved | What you get |
|---|---|---|---|
| the-deal-gauge | On a deal you'd call 'probably fine' | 20-30 min per deal | 10-dimension score + confirmed vs estimated |
| the-pipeline-scanner | Week before a forecast call | 1-2 hrs per review | The deals that are lying to you, ranked |
| the-90-second-brief | Right before a first or re-engagement call | 20-30 min per call | What changed + strongest opening angle |
| the-call-coach | Before and after calls you care about | 15-20 min per call | Prep questions or a scored review + drills |
| the-account-blueprint | Stuck at a single champion | 1-2 hrs per account | Committee map + phased 90-day plan |
| the-objection-playbook | New persona, or same objection keeps losing | 30-45 min per persona | 5 objections with acknowledge/respond/ask |
| the-switch-angle | Competitive displacement, incumbent named | 20-30 min per deal | Questions that surface incumbent pain |
| the-win-loss-analyzer | Quarterly, across 20-40 closed deals | 1-2 days per pass | Evidence-backed ranked reasons |
| the-renewal-tracker | Monthly, on renewals 90-180 days out | 15-20 min per account | Risk score + single highest-impact action |
| the-negotiation-coach | Before any pricing or terms call | 30-45 min per deal | Anchor + concession ladder + walk-away |
Free Intempt tools that pair with these skills
Browser-based, sit behind a free-plan signup, produce quick outputs you can feed into the skills above.
- Meeting notes to CRM: Turn a raw transcript into structured CRM fields - stakeholder, next step, objection - the shape the-deal-gauge and the-call-coach both read.
- Objection response generator: A first-pass response to a specific objection you already have; feed the winners back into the-objection-playbook as your team's actual bank.
- Account research generator: A quick account snapshot to feed into the-90-second-brief when you're prepping five accounts in a row.
What to look for in a Claude skill for sales
Not every prompt is worth wrapping in a skill file. The ones that earn their install share four traits.
Fixed input format. The skill tells you exactly what to paste - stage, close date, transcript, contract value - so you're not deciding structure every time. A skill that takes "whatever you've got" produces whatever it feels like producing.
Consistent output shape. Every run of the-deal-gauge returns a ten-dimension score with a confirmed-versus-estimated split, in the same layout. That's the whole point - a rep can read it in 60 seconds because they know the shape.
Named ceiling. A good sales skill tells you what it can't do. the-renewal-tracker reads one account at a time; the-win-loss-analyzer names patterns but doesn't fix them. Skills that promise everything solve nothing.
Evidence over vibes. The output should point to the transcript line or the usage trend the read is built on, not just assert a conclusion. A 78 with sources reads differently than a 78 that appears from nowhere.
Final thoughts: which skill to install first
Start with these three, in order:
- the-deal-gauge: on the deal you'd call "probably fine." The estimated-versus-confirmed split is the read your CRM was supposed to give you and never did.
- the-90-second-brief: on your next first call. It compresses 30 minutes of tab-hopping into the two lines you actually open with.
- the-objection-playbook: on the objection your team keeps losing on. If the response has been getting invented in real time, that's the one to write down.
For the full AE workflow this pack was built into, see gtm skills for account executives. For the wider catalog, browse the gtm-skills directory. And if you'd rather run the same jobs against live customer data - deal scores that update as usage moves, renewal risk that refreshes daily, objection patterns pulled from every call automatically - start for free on Intempt.
Frequently asked questions. Answered.
Claude skills are versioned SKILL.md files with a fixed input format and a defined output shape, packaged so Claude Code reads them from a local folder. For an account executive, they score deal health and buyer intent, produce a 90-second pre-call brief on any account, prep the five objections a persona actually raises, coach the specific negotiation in front of you, and score a renewal from usage and relationship signals. Install once with npx skills add sidchaudhary/gtm-skills/skills/account-executive and Claude Code picks them up automatically.
