hypothesis-builder
Use when asked to generate campaign hypotheses, build a testing matrix, or create a list of ideas to run from an ICP and signal set
How do I install this agent skill?
npx skills add https://github.com/victor-shulga/outbound-engine-skills --skill hypothesis-builderIs this agent skill safe to install?
- Gen Agent Trust Hubpass
The skill is a marketing tool designed to generate campaign hypotheses from customer profiles and sales signals. It provides a structured framework for prioritizing marketing ideas and specific formatting rules for clear output. No security issues or malicious patterns were detected.
- Socketpass
No alerts
- Snykwarn
Risk: MEDIUM · 1 issue
What does this agent skill do?
Hypothesis Builder
Take an ICP, a set of buyer personas, and a list of signals — generate a full testing matrix of campaign hypotheses to run.
Required tools
Install these before running the skill. None of them is strictly required, the skill also runs on pasted input, but without them the result gets noticeably worse: hypotheses are ranked on wording alone, without checking how many companies carry each signal or whether a real example of it exists.
| Tool | Used for | Required? | Without it |
|---|---|---|---|
| AI Ark, Icypeas or MoltSets (MCP) | company and people counts per persona × signal, to see which hypotheses have enough market to test | Recommended | counts are estimates; a priority-1 hypothesis may have too few accounts to test |
| Sales Navigator | persona counts by title and seniority inside the ICP | Recommended | persona authority is judged from titles on paper |
| Firecrawl (MCP) | reading the company's site and case studies for proof behind each angle | Recommended | angles rest on what you paste; a social-proof angle may have no real case behind it |
| Apify | pulling 3-5 live examples per signal (job posts, LinkedIn posts) before generating | Recommended | vague signals stay vague; you collect examples by hand |
Alternatives: Apollo can stand in for these databases; Exa (MCP) can stand in for Firecrawl.
Inputs the skill expects:
- ICP definition (industry, company size, geography)
- Target personas (2–5 roles you want to reach)
- Buyer signals available (hiring, funding, tool change, leadership change, etc.)
- Angles available (problem-based, time-based, social proof, competitive, etc.)
Process
- Map every combination of: persona × signal × angle
- Remove combinations that are logically incoherent (e.g., "new CRO" signal sent to an SDR)
- Group into clusters by similarity — don't list near-duplicates as separate hypotheses
- Assign a priority score to each hypothesis (1–3) based on: signal specificity, persona authority, angle strength
- Return the full matrix sorted by priority, with a one-line description of each
Terms and plain-language output
<!-- glossary:start terms=hypothesis-matrix,buying-signal,angle,decision-maker -->Rules for every document a client or their team will read (review, roast, canvas, offer cards, instruction):
- Open with «Словник цього документа»: only the terms that actually appear in it, at most 10, each as «термін: що це простими словами». Like the definitions section of a contract.
- No internal codes or letter scales in the text (A/B/C, D L T E, BLEND numbers without words, Own/Shared/Generic, Z1.4, GATE 2.5, T1–T6). Write the meaning instead: «є цифра з клієнтом», not «A».
- Use the plain form from the table below. English stays only for tool names and for copy that ships in English (promises, hooks, subject lines). If the document is in another language, translate the plain form, never the internal term.
- Do not score what cannot exist yet. A block that depends on work not done (signal catalog, case studies, a previous step) gets «пропущено: спершу потрібно <що саме>», not a score.
- Every section of a review opens with one concrete action («до наступного дзвінка: …»), then at most five lines of why. No paragraph that only restates the framework.
| Internal term (this skill) | Write for the reader | What it means |
|---|---|---|
| hypothesis matrix | таблиця гіпотез | список варіантів «кому · з чим · через який канал» для перевірки в розсилках. |
| buying signal | ознака готовності купувати | факт про компанію, після якого розмова про вашу послугу доречна саме зараз. |
| angle, differentiation angle | кут подачі | одна думка, з якою ви заходите до покупця: його біль, ваш доказ і фраза, яку він повторить. |
| decision maker | ОПР (особа, що приймає рішення) | той, хто підписує: власник, операційний директор, керівник напряму |
Output format
Hypothesis matrix: [ICP name]
Generated: [date]
Total hypotheses: [n]
PRIORITY 1 — Highest expected performance
H01 | [Persona] | [Signal] | [Angle] | [One-line description]
H02 | ...
PRIORITY 2 — Worth testing
H03 | ...
PRIORITY 3 — Low conviction, test last
H08 | ...
Recommended starting set: [H01, H02, H04] — run these first
Notes
- Aim for 20–50 hypotheses for a single ICP/vertical — too few means under-testing, too many means you lack focus
- Priority 1 hypotheses should have a specific, recent signal + a senior decision-maker persona
- If signals list is vague, ask for specific examples before generating
- The matrix is not a campaign list — it's a backlog. Only run what you can actually resource
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/victor-shulga/outbound-engine-skills/hypothesis-builder">View hypothesis-builder on skillZs</a>