Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
This skill makes every significant code modification a chance to build genuine understanding. After creating new files, altering schemas, or refactoring, it offers a short exercise that pauses for the developer's prediction, then compares it to the actual behavior. The approach addresses the AI productivity trap—fast output without deep learning. An accompanying orient skill generates a repo-specific guide, so new team members can quickly build a mental model of the codebase. The pair helps AI-assisted developers, new hires, and tech leads turn shipping into deliberate skill development.
WHO IT'S FOR
AI-Assisted Coding Developers
develop skills during AI-assisted coding
Technical Leads / Engineering Managers
guide team skill growth with AI tools
Onboarding Mentors / Codebase Guides
create orientation resources for new members
Developer Education / Training Specialists
design deliberate practice exercises for developers
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Follow the documented setup, then try a first task.
I’m using
Register the GitHub marketplace in Codex
Installs: Registers the complete learning-opportunities Codex plugin marketplace, documented as containing learning-opportunities, learning-opportunities-auto, and orient. The excerpt does not document a subsequent Codex plugin-install command. · Codex
Before you start
Codex with plugin marketplace support
Run this terminal command to add the marketplace from GitHub.
Installs: Registers a local development checkout of the complete learning-opportunities Codex plugin marketplace, documented as containing learning-opportunities, learning-opportunities-auto, and orient. The excerpt does not document a subsequent Codex plugin-install command. · Codex
Before you start
Codex with plugin marketplace support
A local checkout of the learning-opportunities repository
In this terminal command, replace `/path/to/learning-opportunities` with the path to the local repository checkout.
Installs: Installs learning-opportunities-auto, the companion plugin that detects successful git commits and prompts the core learning-opportunities skill to consider offering an exercise. · Claude Code
Before you start
The learning-opportunities core plugin must already be installed
On native Windows, a Git for Windows bash installation is required and Claude Code must be configured to find it
Platform: Linux, macOS, and WSL2 work out of the box. Native Windows requires configuration of the Git for Windows bash path.
After confirming that the learning-opportunities core plugin is installed, enter this command in a Claude Code session.
Enter this command in the Claude Code session to reload plugins.
/plugin reload
On native Windows only, manually set the `CLAUDE_CODE_GIT_BASH_PATH` environment variable to the actual path of the Git for Windows bash executable. The documented typical value is `C:\Program Files\Git\bin\bash.exe`; it may be set as a system environment variable or in the shell profile before launching Claude Code.
A recent new file, module, schema change, architectural decision, refactor, or unfamiliar pattern
The relevant project path or a description of the change
The code or design topic the user wants to understand
Use the learning-opportunities skill to offer me a quick 10–15 minute learning exercise about the recent architectural change in `[path or change description]`. Focus on `[topic I want to understand]`, pause for my answers, and adjust the difficulty based on my responses.
Use the orient skill to generate a repo-specific orientation.md for the repository in `[project working directory]`. Use the project-level destination appropriate for my current agent, create the target directory if needed, and leave its other existing files untouched.
Claude
Codex