One-off prompts don't scale. Every time someone on your team retypes the same instructions into Claude, that's institutional knowledge staying locked in one person's head instead of running as infrastructure.

This is a build session. Daniel builds Claude skills live during the workshop, working from real acquisition criteria, reporting formats, and recurring analyses, then walks attendees through building their own in parallel. Nobody leaves with theory. Everyone leaves with skills that already run.

What you'll build

• Two to three working Claude skills, built live during the session

• A packaged acquisition-criteria skill your whole team can run consistently

• A standardized reporting skill, tested on a real deal or report during the session

• A skill for a recurring analysis, running the same regardless of who's at the keyboard

• A prioritized plan for the rest of your skill library, sequencing what to build next

Hosted by

Daniel Kronovet, CTO, Thesis Driven

You'll learn how to

• Build three Claude skills live during the session, tested against actual workflows

• Identify which of your firm's workflows are actually worth turning into skills

• Structure a skill so it captures the judgment behind a workflow, not just the steps

• Leave with a prioritized plan for the rest of your skill library

Questions about the workshop?
View FAQs on format, access, recordings and more: FAQs

How Claude Skills Actually Work

The mechanics behind the leverage:

• What a skill is and how it differs from a saved prompt

• How Claude uses a skill file to structure its output

• Where skills fit relative to raw prompting and custom tooling

Live Build: Packaging Acquisition Criteria

Daniel builds a criteria skill live, then attendees build their own:

• Identifying the criteria your team applies inconsistently today

• Structuring criteria so Claude applies them the same way every time

• Handling edge cases without breaking the skill, tested in real time

Live Build: Standardizing a Reporting Format

Turning your best format into the default format, built during the session:

• Capturing structure, tone, and level of detail in a single skill

• Applying the skill across different deal types or asset classes

• Attendees build against their own firm's reporting template

Live Build: Encoding a Standard Analysis

Getting a recurring analysis running on demand, live:

• Identifying which analyses get rebuilt from scratch too often

• Structuring inputs and outputs so the analysis runs reliably

• Testing the skill against a real example before the session ends