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 repository collects over thirty skills that give an AI agent specialized abilities across marketing, design, development, and operations. Instead of a single-purpose tool, it offers a curated library: lead research, competitive ad analysis, content writing, algorithmic art, frontend artifact building, meeting analysis, Notion integration, and document processing. It serves professionals who want to extend what an agent can handle without building each capability from scratch.
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.
A local Git repository containing the relevant commit history
The repository root as the working directory
Optional brand or changelog guidelines
Check the repository for this skill’s setup.
Use the changelog-generator skill from the root of my Git repository to draft a user-facing changelog for commits from the past week. Categorize meaningful changes, translate technical commits into customer-friendly language, and filter internal noise. Return the draft for review only; do not publish it or write it to a file.
Path to a local screenshot; replace [path/to/screenshot.png] in the prompt
Intended use: documentation
Check the repository for this skill’s setup.
Use the image-enhancer skill to analyze and improve [path/to/screenshot.png] for documentation. Replace [path/to/screenshot.png] with the actual local image path. Improve sharpness and clarity, reduce compression artifacts if present, preserve the original, and report the enhanced file's path.
One expected behavior to verify; replace [expected behavior] in the prompt
A working Python Playwright environment
Check the repository for this skill’s setup.
Use the webapp-testing skill to verify this behavior in my running local application: [expected behavior]. Open [local URL], wait for the dynamic page to finish loading, inspect the rendered state, perform only the interaction needed for this check, capture a screenshot, and report whether the behavior passed along with any browser-console errors. Replace [expected behavior] and [local URL] with the actual test case and local address. Do not modify application files.
Claude
Codex
Cursor
Gemini CLI
Copilot
OpenCode
Qwen Code