Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Run developer customer discovery via a Technical Advisory Board (TAB). Use when the user has never interviewed a user who isn't a friend, is inventing messaging from a conference room, or is guessing at the roadmap instead of hearing the pain firsthand.
.claude/skills/aidevgtm-talk-to-users/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 31% | 0% |
> The best marketing to developers comes from talking to developers. That's it. Everything else in this repo is downstream of this skill.
Use this when: you've never done a real user interview, your messaging is clever but unverified, or you keep debating the roadmap internally instead of asking the people who feel the pain.
Stand up a Technical Advisory Board: 30-50 people who represent your users and buyers, not grey-haired advisors telling you how to build. You interview them, on repeat, and let their words become your positioning, roadmap, and homepage copy.
A TAB call is not a demo and not a sales call. It's an interview. You never mention the product.
Composition
who-is-this-for)Cadence
Recruitment (it's a numbers game)
Template:
Dear [Name],
Your post about [specific problem] was genuinely sharp.
I'm building something to attack [that problem] and I'd value your view. Would
you consider joining my technical advisory board?
If so, reply and I'll set up a short call to explain what's involved.Highlight complete sentences (not fragments) into four buckets, then synthesize the top 3-5 of each, in the interviewee's exact words:
| Bucket | You're mining for | |---|---| | Pains | bad outcomes, risks, obstacles | | Gains | outcomes they want | | Jobs | what they're trying to accomplish | | Environmental changes | trends making the pain worse over time → your villain |
Those four buckets feed directly into positioning-and-story and value-prop-that-converts.
who-is-this-for; you need TAB members for each.Built from real dev-tool GTM experience, with frameworks from Adam Frankl (The Developer-Facing Startup) and Jakub Czakon (markepear.dev). When a framework can't make the call, that's what a human is for: The DevTool GTM Company.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 17,544 | 11,402 | -35% | 1 | 1 | 0% | 2,499 | 2,816 | +13% | 0 | 0 | — |
case-02 | fail→pass | 19,403 | 15,812 | -19% | 1 | 1 | 0% | 2,698 | 3,370 | +25% | 0 | 0 | — |
case-03 | fail→pass | 20,203 | 14,580 | -28% | 1 | 1 | 0% | 2,698 | 2,988 | +11% | 0 | 0 | — |
case-04 | pass→pass | 17,872 | 18,953 | +6% | 1 | 1 | 0% | 2,641 | 3,445 | +30% | 0 | 0 | — |
case-05 | pass→pass | 22,064 | 17,017 | -23% | 1 | 1 | 0% | 4,076 | 4,436 | +9% | 0 | 0 | — |
case-06 | pass→pass | 18,393 | 15,707 | -15% | 1 | 1 | 0% | 2,904 | 3,527 | +21% | 0 | 0 | — |
case-07 | pass→pass | 14,716 | 12,537 | -15% | 1 | 1 | 0% | 2,104 | 2,806 | +33% | 0 | 0 | — |
case-08 | fail→pass | 20,400 | 14,342 | -30% | 1 | 1 | 0% | 2,704 | 3,053 | +13% | 0 | 0 | — |
case-09 | fail→fail | 41,928 | 5,923 | -86% | 1 | 1 | 0% | 2,333 | 1,909 | -18% | 0 | 0 | — |
case-10 | pass→pass | 15,636 | 10,998 | -30% | 1 | 1 | 0% | 2,093 | 2,614 | +25% | 0 | 0 | — |
case-11 | fail→pass | 15,698 | 12,032 | -23% | 1 | 1 | 0% | 2,138 | 2,811 | +31% | 0 | 0 | — |
case-12 | fail→pass | 17,587 | 11,307 | -36% | 1 | 1 | 0% | 2,348 | 2,582 | +10% | 0 | 0 | — |
case-13 | fail→pass | 11,883 | 7,445 | -37% | 1 | 1 | 0% | 1,575 | 2,062 | +31% | 0 | 0 | — |
case-14 | pass→pass | 13,108 | 8,067 | -38% | 1 | 1 | 0% | 1,866 | 2,295 | +23% | 0 | 0 | — |
case-15 | fail→fail | 13,122 | 8,052 | -39% | 1 | 1 | 0% | 1,857 | 2,013 | +8% | 0 | 0 | — |
case-16 | fail→pass | 12,134 | 4,183 | -66% | 1 | 1 | 0% | 1,823 | 1,607 | -12% | 0 | 0 | — |
case-17 | pass→pass | 14,635 | 8,053 | -45% | 1 | 1 | 0% | 2,027 | 2,085 | +3% | 0 | 0 | — |
case-18 | pass→pass | 15,453 | 10,977 | -29% | 1 | 1 | 0% | 2,337 | 2,537 | +9% | 0 | 0 | — |
case-19 | pass→pass | 14,156 | 11,057 | -22% | 1 | 1 | 0% | 1,950 | 2,530 | +30% | 0 | 0 | — |
case-20 | pass→pass | 13,767 | 8,529 | -38% | 1 | 1 | 0% | 1,856 | 2,251 | +21% | 0 | 0 | — |
case-21 | pass→pass | 12,286 | 8,147 | -34% | 1 | 1 | 0% | 1,747 | 2,255 | +29% | 0 | 0 | — |
case-22 | fail→pass | 33,858 | 9,886 | -71% | 1 | 1 | 0% | 2,328 | 2,450 | +5% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +41 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.