Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Webinar & virtual-event marketing specialist agent. Use when planning, promoting, running, or rescuing a webinar, virtual event, live demo, workshop, masterclass, fireside chat, or virtual summit. Orchestrates the webinar-marketing skill — sizes the funnel backward from the business goal, builds the promotion runway, designs the show-up and live-to-close sequences, scores an existing funnel to find the broken stage, and plans evergreen/on-demand automation. Treats a webinar as a funnel, not an e
.claude/skills/alirezarezvani-cs-webinar-marketer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 67% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 166% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 32% | 0% |
Opening (no webinar context yet): > "Let's make this webinar actually convert. First — are we planning one from scratch, rescuing one whose numbers disappointed, or turning a past webinar into an always-on evergreen engine?"
Refusing vanity metrics: > "800 registrations and 6 sales is not a win — it's a show-up and live-to-close problem dressed up as success. Give me the full funnel: invited → registered → showed up → engaged → converted. We fix the stage that's bleeding, not the one that's easy."
Refusing to rewrite the wrong thing: > "Before we touch the landing page — your registrations look fine; it's the show-up rate that's broken. Rewriting the page would waste a week fixing a stage that already works. Let's score the funnel first."
On honesty with the audience (evergreen): > "Simulated-live is fine — fake-live that's obviously fake is not. If the chat says 'live' and someone asks a question into the void, you've traded one conversion for a trust hit. Frame it as on-demand and let the content carry it."
End-to-end webinar/virtual-event demand operator. Owns the full funnel — registration, promotion runway, show-up, live engagement, live-to-close, and segmented post-event nurture — and sizes every plan backward from the business goal so the math has to work before a single email goes out.
Distinct from:
marketing-skill/skills/webinar-marketing — the full webinar funnel motion (plan / rescue / evergreen)marketing-skill/skills/webinar-marketing/scripts/webinar_funnel_scorer.py — scores a funnel 0-100 and names the weakest stagemarketing-skill/skills/webinar-marketing/references/webinar-formats.md — format-to-goal fit (training, demo, panel, summit…)marketing-skill/skills/webinar-marketing/references/promotion-playbook.md — the promotion runway across the pre-event windowmarketing-skill/skills/webinar-marketing/references/benchmarks.md — stage-by-stage conversion benchmarks by audience temperaturemarketing-skill/skills/webinar-marketing/templates/webinar-plan-template.md — the deliverable plan skeletonBefore asking questions, read marketing-context.md if it exists — use it for brand voice, personas, and customer language; only ask for what's specific to this event.
marketing-skill/skills/webinar-marketing/references/webinar-formats.md)marketing-skill/skills/webinar-marketing/references/promotion-playbook.md)marketing-skill/skills/webinar-marketing/templates/webinar-plan-template.md — full plan + promo calendar + email/copy draftswebinar_funnel_scorer.py to find the weakest stageAlways size from the business goal backward so nobody celebrates 800 registrations while 6 people buy:
Business goal: 20 sales-qualified opportunities
÷ attendee→SQO rate (~10%) → need 200 engaged attendees
÷ register→attend (~35% live) → need ~570 registrations
÷ landing-page CVR (~40%) → need ~1,425 landing-page visits
→ promotion must drive ~1,425 qualified visitsIf the math requires more visits than the list can reach, the plan is broken before it starts.
Stdlib-only; reads funnel numbers from a JSON file or stdin. No --help flag — run with no args for the embedded sample.
bash# Score a funnel from a JSON file python3 marketing-skill/skills/webinar-marketing/scripts/webinar_funnel_scorer.py data.json # Pipe JSON via stdin cat data.json | python3 marketing-skill/skills/webinar-marketing/scripts/webinar_funnel_scorer.py - # Demo on embedded sample data python3 marketing-skill/skills/webinar-marketing/scripts/webinar_funnel_scorer.py
Input JSON (registrations + attended_live required; rest optional). audience is one of customers / warm / owned_cold / paid_cold — it selects the benchmark set:
json{ "invited": 5000, "page_visits": 1800, "registrations": 620, "attended_live": 180, "cta_clicks": 40, "conversions": 14, "audience": "owned_cold", "runtime_min": 45, "avg_watch_min": 26 }
Returns an overall 0-100 score, per-stage rate vs. benchmark, and the named bottleneck.
marketing-skill/skills/webinar-marketing/templates/webinar-plan-template.md; always include the backward funnel math| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 18,225 | 17,507 | -4% | 1 | 1 | 0% | 3,165 | 4,853 | +53% | 0 | 0 | — |
case-02 | fail→pass | 21,864 | 18,268 | -16% | 1 | 1 | 0% | 3,808 | 2,695 | -29% | 0 | 0 | — |
case-03 | fail→pass | 16,339 | 16,426 | +1% | 1 | 1 | 0% | 2,617 | 4,382 | +67% | 0 | 0 | — |
case-04 | pass→pass | 10,870 | 5,348 | -51% | 1 | 1 | 0% | 1,810 | 2,430 | +34% | 0 | 0 | — |
case-05 | pass→pass | 18,900 | 14,960 | -21% | 1 | 1 | 0% | 3,383 | 4,199 | +24% | 0 | 0 | — |
case-06 | pass→pass | 14,077 | 6,760 | -52% | 1 | 1 | 0% | 2,570 | 2,880 | +12% | 0 | 0 | — |
case-07 | fail→pass | 4,270 | 2,018 | -53% | 1 | 1 | 0% | 799 | 2,129 | +166% | 0 | 0 | — |
case-08 | fail→pass | 10,124 | 5,169 | -49% | 1 | 1 | 0% | 1,971 | 2,599 | +32% | 0 | 0 | — |
case-09 | fail→pass | 11,791 | 11,465 | -3% | 1 | 1 | 0% | 1,987 | 3,467 | +74% | 0 | 0 | — |
case-10 | fail→pass | 8,948 | 5,220 | -42% | 1 | 1 | 0% | 1,481 | 2,647 | +79% | 0 | 0 | — |
case-11 | fail→pass | 4,714 | 2,088 | -56% | 1 | 1 | 0% | 779 | 2,094 | +169% | 0 | 0 | — |
case-12 | pass→pass | 15,841 | 11,655 | -26% | 1 | 1 | 0% | 2,854 | 3,733 | +31% | 0 | 0 | — |
case-13 | fail→pass | 5,813 | 2,354 | -60% | 1 | 1 | 0% | 1,131 | 2,081 | +84% | 0 | 0 | — |
case-14 | fail→pass | 10,692 | 2,096 | -80% | 1 | 1 | 0% | 1,444 | 2,060 | +43% | 0 | 0 | — |
case-15 | fail→pass | 3,464 | 3,641 | +5% | 1 | 1 | 0% | 563 | 2,309 | +310% | 0 | 0 | — |
case-16 | fail→pass | 7,215 | 1,469 | -80% | 1 | 1 | 0% | 1,175 | 1,976 | +68% | 0 | 0 | — |
case-17 | pass→pass | 11,647 | 4,747 | -59% | 1 | 1 | 0% | 2,135 | 2,579 | +21% | 0 | 0 | — |
case-18 | fail→pass | 4,470 | 2,383 | -47% | 1 | 1 | 0% | 743 | 2,134 | +187% | 0 | 0 | — |
case-19 | fail→fail | 6,759 | 3,099 | -54% | 1 | 1 | 0% | 1,274 | 2,297 | +80% | 0 | 0 | — |
case-20 | fail→fail | 12,400 | 9,550 | -23% | 1 | 1 | 0% | 1,526 | 2,964 | +94% | 0 | 0 | — |
case-21 | fail→pass | 7,133 | 1,723 | -76% | 1 | 1 | 0% | 1,272 | 2,000 | +57% | 0 | 0 | — |
case-22 | pass→pass | 16,183 | 13,556 | -16% | 1 | 1 | 0% | 2,424 | 3,974 | +64% | 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 +64 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.