What happens during the 'Counterpoint' phase?

The Counterpoint phase is a mandatory step where the AI is required to identify its own blind spots and surface alternative viewpoints to ensure a balanced evaluation.

What is a 'Thinking Profile' in this skill?

A Thinking Profile is a specialized configuration that defines an expert's voice, priorities, and lens, allowing Claude to simulate a specific professional's approach to code review.

Where are my review sessions saved?

Review results are automatically logged as Markdown files in your home directory under ~/.things/think-like/sessions/ for easy retrieval.

Does it support reviewing multiple files at once?

Yes, the skill features a parallel execution mode that launches multiple sub-agents to analyze several targets simultaneously, providing a synthesized final report.

Can I use my own custom experts?

Absolutely. You can create and install custom profiles in your local directory, defining your own rules for voice, red flags, and output formats.

Expert Thinking Profiles

Name: Expert Thinking Profiles
Author: brennacodes

bybrennacodes

0•

협업 도구

Apply specialized architectural and coding philosophies to your codebase through the perspective of industry-leading experts.

The Profile skill transforms Claude into a specialized consultant by applying curated 'thinking profiles' to your development workflow. By loading specific action files, Claude can perform code reviews, security audits, and architectural evaluations through the lens of experts like DHH or Sandi Metz. It supports both single-target analysis and high-performance parallel execution for monorepos, ensuring every piece of code is validated against specific design patterns, rhetorical styles, and mandatory counterpoint perspectives.

주요 기능

01Parallel agent execution for analyzing multiple directories simultaneously

02Mandatory 'Counterpoint' analysis to surface hidden architectural blind spots

03Automated session logging for audit trails and historical reference

04Extensible profile system supporting custom voice, priorities, and red flags

050 GitHub stars

06Expert-driven code reviews using industry-standard philosophies

사용 사례

01Running a parallel security audit across multiple microservices or plugins

02Conducting architectural reviews focused on simplicity and Rails-style conventions

03Evaluating object-oriented design patterns using a Sandi Metz profile

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add brennacodes/brenna-plugs profile

For use in Claude.ai and ChatGPT