madogiwa0124

Sends current git changes to GitHub Copilot CLI for code review and displays the results. Use when requesting a second-opinion review from Copilot on staged changes, uncommitted work, or a branch diff.

Generates a structured Git commit message (header + Motivation / Overview / Main Changes / Breaking Changes sections) from staged diffs. Use when the user asks to write or generate a commit message.

Supports the creation and improvement of prompts for various use cases.

Get the current date and time. Use when a log timestamp or time check is needed.

Provides guidelines and best practices for creating and improving new skills for agents. (ref: https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices)

Guide for evaluating web content against WCAG 2.1 Level A accessibility standards. Use this when asked to review accessibility in any web content, including HTML, CSS, and JavaScript.

Guidelines for writing commit messages in English using the Conventional Commits format. Use this when asked to write or review commit messages for the project.

Guide for reviewing modern CSS framework. Use this when asked to review CSS in `packages/css`.

Guide for developing Storybook stories for CSS components in `@madogiwa-ui/css`.

Generates a structured Git commit message (header + Motivation / Overview / Main Changes / Breaking Changes sections) from staged diffs. Use proactively whenever creating a git commit — do not write commit messages manually. Also use when the user explicitly asks to write or generate a commit message.

Run an evidence-based audit of this repository and produce a scored evaluation report covering architecture, code quality, public API design, testing, security, robustness, documentation, packaging, and performance. Use when asked to evaluate the repository's quality, find improvement points, or assess release readiness.

Run the smoke test against local Ollama to verify core features work with a real LLM. Invoke proactively (without being asked) after completing any change that touches: LLM adapter streaming or tool call accumulation, Agent think/llm_call loop, EventListener, Session#chat or Session#live, Strategy execution, or LLM::Client routing. These are areas where dummy-based unit tests pass but real LLM behavior can diverge.

Releases a new patch version of vite-plugin-rollbar-sourcemap to npm. Bumps version in package.json, commits, tags, pushes the tag, and publishes.

Generates OSS-quality Git commit messages that double as PR descriptions. Use when you want a subject line + body ready for both git log and GitHub Pull Requests.