skills.rest
elvisun
elvisunCOMMUNITY
Explore all skills by this author →
Website𝕏@elvissunGitHub

icp-evidence-analysis

Convert source-bound company evidence into testable ideal-customer-profile hypotheses with confidence ratings.

663|47|Updated May 19, 2026
Search Tags:#confidence scoring#segmentation#buyer personas#icp#ideal customer profile#evidence analysis#Market Research
One-click install
npx skills add https://github.com/elvisun/newsjack --skill icp-evidence-analysis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: icp-evidence-analysis
Source: https://github.com/elvisun/newsjack/tree/main/skills/icp-evidence-analysis
Command: npx skills add https://github.com/elvisun/newsjack --skill icp-evidence-analysis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Teams often build ideal customer profiles from gut feeling or a company's own marketing copy, producing persona fiction that cannot survive scrutiny. This Skill turns a source-bound company and market dossier into defensible, evidence-traced ICP hypotheses with explicit confidence levels, counterevidence, and open research questions.

Core Features & Use Cases

  • Evidence Perimeter Building: Classifies every material claim as company_asserted, buyer_behavior, or independent, retaining source IDs, URLs, dates, and permission status while preserving negative and conflicting evidence.
  • Testable ICP Hypotheses: Generates context-based hypotheses covering triggers, buying roles (champion, economic buyer, blocker), constraints, disqualifiers, standing, and counterevidence — never invented demographics.
  • Structured Handoff: Outputs a Markdown summary plus an icp_hypotheses.json artifact with a Gate 1 decision (ready_for_human_review, needs_research, or stop_permission_failure) for downstream buyer-job-intent-analysis.
  • Use Case: Given a startup's URL and a monitor profile, produce low-to-high confidence ICP hypotheses that separate what the company claims from what independent buyer behavior actually supports, before building an AI-visibility prompt panel.

Quick Start

Analyze this company URL and evidence manifest to produce evidence-bound ICP hypotheses with confidence ratings and a Gate 1 decision.

Frequently Asked Questions about icp-evidence-analysis

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build an ideal customer profile from evidence instead of assumptions?▼

Collect public product, pricing, review, and coverage sources, classify each claim by provenance, then form hypotheses with triggers, buying roles, and constraints. Every field must trace to a source ID or remain null, and website-only ICPs stay low confidence until independently supported.

What inputs does ICP evidence analysis accept?▼

It accepts a company URL with description, a source_manifest.json, public product, pricing, security, filing, review, or coverage pages, an existing monitor profile as unverified leads, and target markets or exclusion rules. A URL alone triggers independent evidence research.

How are confidence levels assigned to ICP hypotheses?▼

High requires multiple agreeing sources including direct behavior or strong independent evidence. Medium means one strong source or several consistent weaker ones with gaps. Low applies when only company assertions, sparse proxies, or conflicting evidence exist.

Can ICP analysis use a company's own marketing claims as evidence?▼

Yes, but target-authored content is always classified as company_asserted, never independent. A core-panel ICP needs at least one independent or behavioral source; website-only hypotheses remain hypothesis_only unless a human explicitly promotes them.

What happens after the ICP hypotheses are generated?▼

The skill outputs a Markdown summary and icp_hypotheses.json with a Gate 1 decision. After human review, approved ICP IDs, permitted sources, and open questions are handed to buyer-job-intent-analysis; hypotheses are never silently promoted.

Related Skills

Explore More Marketing & Sales Skills→
a3lema3lem

frontmatter

Stamp and scan AI provenance metadata blocks in source files.

Community
Intermediate
ToniIAPro73ToniIAPro73

nexus-buyers-v1

Automate buyer intake scoring and prospecting visibility across CRM referrals and inbound strategies.

Community
Advanced
bitwize-music-studiobitwize-music-studio
408

researcher

Gather and cross-verify primary-source evidence across 3+ independent sources.

Official
Advanced
matteotittamatteotitta
29

customer-interviews

Generate structured interview guides and synthesis frameworks for B2B SaaS research.

Community
Intermediate