BINGLEY
Sully Fazie

Cold Email Builder

5 mins

Paste any cold email. Get it scored out of 100 and rebuilt on real reply data. Not a random AI generator.

QUICK START
Claude Cowork already set up?

If you've already got Claude Cowork installed and you're on a paid plan, skip the setup and just grab the skill.

Download the skill
First time here?

The full setup guide is right below. It only takes a few minutes.

Before you start

If you don't have the Claude desktop app, you'll need it first, so download it here.

(You'll also need Claude Pro, around $20.)

Part 1 · Claude setup(3 mins)
1

Install Claude

Download the Claude desktop app and sign in on a paid plan. It has to be the desktop app, not the browser, because that is the version that can run skills.

2

Turn on code execution

Click your name bottom-left, then Settings > Capabilities, and switch on Code execution and file creation.

Part 2 · Cold Email Builder setup(2 mins)
1

Install the skill

Download the skill, then in Customize > Skills click +, upload the file, and make sure its toggle is on.

Download the skill
2

Start it

Open any new chat and type /cold-email-builder.

Or click + at the bottom of the chat, choose Skills, and pick Cold Email Builder.

Then paste the cold email you want scored. A screenshot is fine.

Before you finish (quick checks)

New How was this skill built? (Click here)

How was this skill built?

Every skill ships only after planning, research, building, testing and re-testing. Here is the record for the builder.

Built on 190M+ real emails Peer-reviewed + platform-official sources 94% inter-rater agreement
PlanningOne rubric used both directions, to build and to grade. A grade-first front door, and it scores the email, never the person.
Research & background checksBuilt on studies covering more than 190 million real cold emails across the major datasets (Yesware, Backlinko, Gong, Hunter), plus peer-reviewed field experiments and the official sender rules from Google, Yahoo, Microsoft and Apple. Every rubric weight traces back to a cited source.
BuildA live scorecard artefact that scores against the 7 things that actually drive replies, with every judgement evidence-tagged to a source.
Testing & re-testingStress-tested across 10 runs, blind inter-rater reliability 94% within one band. An outcome harness keeps learning from real sent-email results.
OrganisationVersioned rubric, plain-language layer, and evidence base kept in step.

    Key strengths

  • Grounded in studies of 190M+ real emails, not a random AI generator.
  • Every judgement evidence-tagged with its source.
  • Builds and grades on one calibrated rubric.
Want access to all of the Bingley modules?
You'll find it in my free community full of people in sales working with AI.
Join free
Sully Fazie
By Sully Fazie
sullyfazie.com