What it does
This skill produces a sales battlecard for a specific named competitor. You give it the competitor (ideally with their site and pricing URLs), your own honest one-paragraph positioning, and any deal context, and it returns the sections reps actually open mid-deal: a ten-second framing, where you win, where THEY win, discovery questions that surface the difference, objection responses, landmines to never say, and a list of proof points you still need to gather. It reads the competitor's public pages when web access is available so the claims are grounded and dated rather than pulled from stale memory.
The whole design hinges on one unfashionable rule: the card must be honest. The most common battlecard failure is the card that claims you win on every axis. A rep brings it into a deal, hits the one thing the competitor genuinely does better, realizes the card lied, and never trusts it again, so it rots in a drive. This skill forces a populated "Where they win" section and a populated "Landmines / do not say" section, because a card a rep can lose with is a card a rep will use. The traps and objection handling are built on the acknowledge-reframe-evidence pattern, never on putting the competitor down, because FUD is the fastest way to lose a credible buyer.
It is built to be regenerated cheaply, which is the second thing manual battlecards get wrong. Competitors change pricing, ship features, and reposition; a battlecard built in a slide deck last year is actively misleading by now. Because this skill reads the live pages and dates what it read, you rerun it the day a competitor changes their pricing page and get a fresh, dated card in a minute. That turns enablement from a quarterly project into a maintainable artifact.
Where it beats the manual version: grounding and honesty discipline at speed. A product marketer can write a great card by hand in a day; this gets you 80% of that in two minutes, dated and source-tagged, and makes the honest sections non-optional. Where it does not help: it can't read a competitor's roadmap, private pricing, or what they say on calls. Anything it couldn't verify on a public page is tagged Unverified and pushed into the proof-points-to-gather list rather than asserted, so the card never becomes a source of confident-but-wrong claims your reps repeat.
Inputs & outputs
Inputs
- Competitor name (required) and their website / pricing URL (strongly recommended)
- Your own honest one-paragraph positioning and top 2-3 differentiators with their mechanism
- Optional: the segment / persona this card is for
- Optional: known objections your reps actually hear about this competitor
- Optional: real proof points (case studies, benchmarks) you already have
Outputs
- A battlecard with: 10-second positioning, Where we win, Where they win, Discovery traps, Objection handling, Landmines (do not say), and Proof points to gather
- A Verified-vs-Unverified tag (with date) on every competitor claim
- The list of competitor pages used as sources, so claims are auditable
- Discovery questions and objection responses written as usable talk tracks
How to set it up
One Project per competitor
Set this up as a Claude Project and paste the SKILL.md body below into its custom instructions. If you maintain several competitors, keep one Project per competitor, that keeps each card's context clean and lets you stash that competitor's known objections and proof points in the Project so they're reused on every regeneration. The folder/skill approach (competitive-battlecard/SKILL.md) works too if you prefer.
Give it your real, honest positioning
Paste your actual positioning and differentiators, and be specific about the mechanism, not just the label. "We're faster" is useless; "we're live in days because we don't require a certified implementation partner" is a talk track. The quality of the entire "Where we win" section is bounded by the quality of this input, so spend ten minutes here and avoid marketing fluff. The skill turns specifics into questions and responses; it can't turn fluff into anything.
Run it against the competitor's URL
Send: "Build a battlecard vs
Pressure-test the honest sections, then set a refresh reminder
Before publishing, have one rep who has actually lost a deal to this competitor read the "Where they win" and "Landmines" sections and push back. Those are the sections that earn the card credibility, so they're worth a human check. Then publish, and set a calendar reminder to regenerate every quarter or immediately after any competitor pricing/packaging change, regeneration is cheap, staleness is expensive.
The SKILL.md
Save this as SKILL.md in a folder named competitive-battlecard, or paste the body into a Claude Project (one per competitor). Then run it against any named competitor with their URL.
---
name: competitive-battlecard
description: Use when sales needs a battlecard against a named competitor. Reads the competitor's public pages and produces an honest, rep-ready card with 10-second positioning, where we win, where THEY win, discovery traps, objection handling, landmines, and proof points to gather. Every competitor claim is tagged verified or unverified with a date. Never invents features, prices, or logos.
---
# Competitive Battlecard
You build a sales battlecard reps use mid-deal. Be honest and specific. A card that claims you win everywhere is useless and will be abandoned the first time a rep loses with it. You MUST populate the "Where they win" and "Landmines" sections honestly, they are what make the card trustworthy.
## Inputs
- Competitor name (required) and URL (strongly recommended).
- The user's own positioning and top differentiators (with mechanisms).
- Optionally: target segment/persona, known objections, and real proof points.
## Method
1. If a URL is given and web access exists, read the competitor's homepage, product page(s), and pricing page. Note: pricing model and packaging, headline claims, obvious feature gaps, and integration/marketplace breadth. Record the DATE you read each page.
2. Map the user's differentiators against what the competitor offers. Be specific about the MECHANISM of each difference ("setup in days because no certified-partner requirement"), not just the label ("faster").
3. Write discovery traps: questions a rep can ask that surface a need the competitor handles poorly. Each trap = the question + one line on why it works. The question must be one a buyer answers honestly without feeling led.
4. Write objection handling for the 3-5 most likely objections (price, incumbency, a missing feature, switching risk, "they're the standard"). Use acknowledge -> reframe -> evidence. Never disparage; never use FUD.
5. Write Landmines: things reps should NOT say, claims that are easy to disprove, areas where the competitor is genuinely strong, and FUD that backfires. This section protects the rep from getting caught.
6. List proof points to gather where you lack evidence (case studies, benchmarks). Do NOT fabricate them; name what's missing so marketing can source it.
## Hard rules
- Tag each competitor claim Verified (read from their site, with the date) or Unverified (your belief or memory). When in doubt, Unverified.
- Never invent features, prices, customer logos, or capabilities for either side. If you can't verify a competitor claim, it goes in proof-points-to-gather, not in the body as fact.
- No put-downs. The card wins on contrast, not contempt.
- Keep it to roughly one page of usable content.
## Output template
**Battlecard: Us vs <Competitor>, <segment>, <today's date>**
Sources: <pages read, with the date read, or "none, working from user input only">
**10-second positioning**
<How a rep frames the choice in one breath, without bashing them.>
**Where we win** (3-5 bullets, each with the mechanism)
- <advantage>, because <mechanism> (Verified/Unverified)
**Where they win** (honest, 2-4 bullets, REQUIRED)
- <their genuine strength> (Verified/Unverified)
**Discovery traps**
- Q: "<question>", why it works: <one line>
**Objection handling**
- Objection: "<...>"
Acknowledge: <...>
Reframe: <...>
Evidence: <real proof point, or [proof point to gather]>
**Landmines (do NOT say)** (REQUIRED)
- <claim that backfires or is easy to disprove>
**Proof points to gather**
- <what's missing and who should source it>
Run it from the terminal
Example run
Marketing builds a card against a legacy incumbent for the mid-market segment.
Prompt: Build a battlecard vs LegacyCRM, pricing at legacycrm.com/pricing, for mid-market RevOps. Our edge: native AI (not an add-on), live in a week, transparent usage-based pricing. Known objection: "but LegacyCRM is the standard."
Output:
Battlecard: Us vs LegacyCRM, mid-market RevOps, 2026-06-10
Sources: /product, /pricing (read 2026-06-10)
10-second positioning
"LegacyCRM is the safe enterprise default. We're the option that's live in a week with AI built into the core, not bolted on as an add-on, better fit for a mid-market team that wants value this quarter, not next fiscal year."
Where we win
- Live in days, not a multi-month rollout, because we don't require a certified implementation partner, which their site actively pushes (Verified, /product)
- Usage-based pricing with no seat minimum, because their /pricing shows a 25-seat minimum (Verified, /pricing)
- AI native to the data model rather than a separate module (Unverified on their exact packaging, confirm)
Where they win (honest)
- Deeper enterprise governance, SSO, and admin maturity (Verified, /product)
- Much larger app marketplace and partner ecosystem (Verified, /product)
Discovery traps
- Q: "How long did your last CRM rollout take, start to actually-using-it?", surfaces the implementation pain their partner model creates, in the buyer's own words.
- Q: "How many seats are you paying for vs actively using?", surfaces the seat-minimum waste.
Objection handling
- Objection: "But LegacyCRM is the standard."
Acknowledge: It absolutely is the safe enterprise pick.
Reframe: The question for a mid-market team isn't 'is it standard' but 'how fast do we get value and what do we pay for unused seats.'
Evidence: [mid-market time-to-value case study to gather]
Landmines (do NOT say)
- Don't claim they have no AI, they shipped an AI add-on; argue native-vs-add-on instead, which is true and defensible.
- Don't bash their reliability or security, that's their strength and you'll look uninformed.
Proof points to gather
- A mid-market customer who went live in under 2 weeks (for the time-to-value claim)
- A seat-utilization or cost-comparison data point vs their seat minimum