grill-me
Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree. Use when user wants to stress-test a plan, get grilled on their design, or mentions "grill me".
agent-browser
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.
cosense-memo
Claudeとの会話内容や記事の要約をCosense形式のメモに変換する。箇条書き・だである調・internal linkを用いた知識ネットワーク向けフォーマット。「Cosenseにメモして」「この内容をCosense形式で」「学んだことをまとめて」などのリクエストで使用。Cosenseは旧称ScrapboxのためユーザーがScrapboxと言ったときもこのスキルを使ってください。
find-skills
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
gog
Google Workspace操作（Gmail, Calendar, Drive, Docs, Sheets, Chat, Tasks, Contacts）。ユーザーが「メール確認して」「予定を見せて」「ドライブで検索」「ドキュメント作成」「スプレッドシート読んで」「タスク追加」「連絡先検索」などと言ったときに使用。gog CLIを利用してGoogle Workspaceの各サービスを操作する。
magi-decision-support
Multi-perspective decision support using three AI agents (Scientist, Mother, Realist). Use when facing difficult decisions, trade-offs, comparing alternatives, choosing between options, or when unsure which approach is best. Triggers on questions like "should I", "which is better", "pros and cons", "normalize vs denormalize", "microservices vs monolith", or any technical/architectural decision with multiple valid options.
pr-lead-time
Measures PR lead time (open -> base-branch merge -> deploy) in business hours (JST, weekends/public holidays excluded), aggregated weekly or monthly. Use when the user asks about PR/delivery lead time, "PRリードタイム", "デリバリーのリードタイム", KR1 progress ("開発着手から本番環境に届くまでの摩擦"), or wants to know how long PRs take to open->merge->deploy for a repo.
review
Conduct comprehensive code reviews using 5 specialized subagents (code quality, performance, test coverage, documentation accuracy, security) in parallel, providing an integrated report. Use at important milestones (5+ file changes, public API updates, infrastructure config updates, before commits/PRs/releases). Triggers on requests like "review this", "check the code", "review the changes".
spotify-playlist-creator
>
summarize
記事・YouTube・PodcastなどのURLを要約する。ユーザーがURLを渡して「要約して」「まとめて」「summarize」などと言ったときに使用。summarize CLIを利用してコンテンツを取得・要約する。