ICP Screener — Claude Code Skill

A Claude Code skill that screens interview transcripts against an Ideal Customer Profile (ICP). It reads only the “profiling information” section of each transcript, classifies each as match / no match / uncertain, and copies matches into a temporary folder with a screening overview — ready for downstream analysis (opportunity extraction, insight mining, etc.).

Run this before any per-transcript analysis so the later steps work on a clean, ICP-relevant subset.

What it does

• Reads your ICP from company context, or asks you for it
• Detects whether transcripts are organized by ICP folder, or mixed in one folder
• For mixed folders, screens each transcript by its profiling section
• Creates icp-screened-TEMP-[YYYY-MM-DD]/ inside your transcripts folder
• Copies matched transcripts into that folder
• Writes screening-overview.md with match / no-match / uncertain tables and reasoning

See SKILL.md for the full skill definition.

Install

Claude Code loads skills from ~/.claude/skills//SKILL.md.

git clone https://github.com/Elsevanderberg1/ICP-screen-claude-code-skill.git ~/.claude/skills/icp-screener

Or, if you prefer to copy just the skill file:

mkdir -p ~/.claude/skills/icp-screener
curl -o ~/.claude/skills/icp-screener/SKILL.md \
  https://raw.githubusercontent.com/Elsevanderberg1/ICP-screen-claude-code-skill/main/SKILL.md

Restart Claude Code (or start a new session) to pick up the skill.

Use

/icp-screener 

Claude will ask for your ICP if it isn’t already in context, then screen the transcripts and produce the output folder.

Requirements

• Claude Code
• One or more interview transcripts, each with a “profiling information” section

License

MIT — see LICENSE.