A set of Claude Code Skills that turn your startup profile into a disciplined go-to-market motion for finding the right design partners and early customers — and screening out the ones that quietly kill startups.
The wrong early customer is as dangerous as no customer: the tire-kicker with no real pain, the champion with no budget authority, the whale who hijacks your roadmap, the free-pilot-forever logo that never converts, the vanity name that proves nothing because no one there actually uses the product. ClientMate is the bad-fit filter — and the engine for finding the partners who give you revenue, references, data, and learning.
It is the sibling of VCupid (fundraising) and Incubae (program evaluation), and reads the same STARTUP_PROFILE.md — maintain one profile, use all three toolkits.
Run commands in this order to build and qualify your design-partner pipeline:
# 0. Create your STARTUP_PROFILE.md (shared with VCupid and Incubae)
# 1. Define who you should — and shouldn't — sell to
/cmstrat
# 2. Build the target account list
/cmlist
# 3. For each Tier 1 account — bad-fit check first:
/cmposer <account name> # Will this relationship advance or bleed you? Drop if score < 40.
# 4. For each account that passes (score 60+):
/cmmatch <account name> # Map the deal, the champion, and the path in
/cmvalue <account name> # What do they give you beyond revenue, net of cost to serve?
# 5. Harden your pitch against a real buyer before any call:
/cmdevil
# 6. For each account with a /cmmatch "Pursue" verdict:
/cmreach cmmatch-<account>.md # Draft outreach to the champion
# 7. When a discovery call is booked:
/cmprep cmmatch-<account>.md
File naming convention:
| File | Purpose |
|---|---|
STARTUP_PROFILE.md |
Your startup data — shared across VCupid, Incubae, and ClientMate |
cmstrat.md |
Strategy — your Ideal Design Partner Profile, ICP, asks, anti-patterns |
cmlist.md |
Master target account pipeline, tiered |
cmposer-<account>.md |
Bad-fit check — pain, budget, velocity, roadmap, conversion |
cmmatch-<account>.md |
Deep fit analysis — 7-dimension score, champion map, deal shape |
cmvalue-<account>.md |
Value audit — commercial + strategic, netted against cost to serve |
cmreach-<account>.md |
Outreach to the champion — cold email, warm intro, LinkedIn |
cmprep-<account>.md |
Discovery call prep + disqualify triggers |
cmdevil.md |
The skeptical buyer — 10 deal-ending questions |
These commands install as the clientmate Claude Code plugin (works in any project directory).
1. Clone the plugin:
git clone https://github.com/maxoliverbr/clientmate-plugin.git ~/dev/clientmate
2. Run the install script:
cd ~/dev/clientmate && bash install.sh
The script registers the plugin in ~/.claude/plugins/installed_plugins.json and enables it in ~/.claude/settings.json. It is idempotent — safe to re-run after updates.
3. Restart Claude Code. All /cm* commands become available in any directory that contains a STARTUP_PROFILE.md.
Manual installation
Register in ~/.claude/plugins/installed_plugins.json:
"clientmate@local": [{
"scope": "user",
"installPath": "/home/<you>/dev/clientmate",
"version": "1.0.0",
"installedAt": "<ISO timestamp>",
"lastUpdated": "<ISO timestamp>"
}]
Enable in ~/.claude/settings.json:
"enabledPlugins": {
"clientmate@local": true
}
Start here. Defines the beachhead, a precise Ideal Design Partner Profile and ICP, how many partners to sign and in what order, exactly what to ask each for (with the terms that protect you), and the anti-patterns to refuse — the roadmap hijacker, the no-budget champion, the free-pilot-forever logo. Reads STARTUP_PROFILE.md (+ any cmlist.md, cmmatch-*.md, cmposer-*.md). Saves cmstrat.md.
Researches the beachhead and produces a tiered list of 15–25 candidate accounts ranked by pain acuity, buyability, reference value, reachability, and beachhead fit — each with a likely champion and an entry angle. Saves cmlist.md.
The core filter. Eight evidence-based checks — Pain Acuity, Budget & Authority, Decision Velocity, Roadmap Alignment, Conversion Intent, Reference Value, Engagement Capacity, Counterparty Health — scored 0–100. Verdicts: Strong Fit (80–100), Workable (60–79), Caution (40–59), Walk Away (0–39). A dream logo with low pain can be a Walk Away. Saves cmposer-<account>.md.
cmposer ≠ cmmatch. Bad-fit Score answers "will this relationship advance or bleed us?" cmmatch answers "what's the deal, who's the champion, what do we propose?" You need both.
Deep research on a single account against your profile. Seven-dimension fit score (Pain–Solution 25, Buyability 20, Reference & Strategic Value 15, Champion Strength 15, Roadmap Fit 10, Data/Domain Access 10, Reachability 5), a champion map, the deal shape to propose (with a mandatory written success definition and conversion path), and a Pursue / Nurture / Pass verdict. Saves cmmatch-<account>.md.
What a partner is worth beyond the contract. Part A (commercial: contract, expansion, reference economics) and Part B (strategic: data & domain access, co-development, distribution, brand signal, learning) — then nets value against cost to serve (support burden, customization, discounts, founder time) for a verdict and a maximum-justified-concession number. The best design partner is often not the biggest payer. Saves cmvalue-<account>.md.
Turns a cmmatch report into three outreach variants aimed at the champion — cold email, warm intro request, LinkedIn DM — each built on the account's sourced, specific pain with a low-friction first ask (a short call, never a contract). Stops if the verdict is Pass. Saves cmreach-<account>.md.
A timed agenda, the qualifying questions that confirm pain/budget/authority/timeline/success-criteria, a brief value story, objection handling (including "we could build this ourselves"), the deal shape and ask, and the disqualify triggers that tell you to walk. A fast, polite no protects your runway. Saves cmprep-<account>.md.
A burned-before enterprise VP asks the 10 hardest deal-ending questions — vendor survival, build-vs-buy, the pilot trap, security, lock-in. Run it before any call; the answers that run out before 30 seconds are the gaps to close. Saves cmdevil.md.
- Keep
STARTUP_PROFILE.mdcurrent. As your ICP sharpens, re-run/cmstrat— the anti-patterns list is the most valuable thing it produces. - Run
/cmposerbefore/cmmatch. An account scoring under 40 is a no — don't invest the relationship time, however shiny the logo. - Never sign a pilot without a written success definition and a conversion path. Every skill in this toolkit enforces this; it's the single most common early-stage trap.
- Weight design-partner value, not just contract size.
/cmvalueexists because the best partner often pays little and teaches a lot. - A fast, polite no is a win.
/cmprep's disqualify triggers are there to save your runway, not to close every deal. - Update the plugin with new skills by adding
SKILL.mdfiles toskills/<name>/and re-runninginstall.sh.
MIT © 2026 3Flux