nanameru

Use when creating, updating, backing up, publishing, restoring, or auditing agent Skills in a GitHub-backed canonical skills repository. Trigger on Skill管理, Skills管理, skills repo, Vercel Skills CLI, npx skills add/update, GitHubにSkillを保存, Skillをpush, restore-skills.sh, ~/.agents/skills backup, prevent skills from disappearing, or when a user wants the management of Skills itself to be handled by a Skill.

Slide design reference workflow: before authoring any slide deck (html2pptx, PPTX, presentation HTML), browse SlideLand (slideland.tech) for real-world reference decks matching the topic, extract design essence (palette/whitespace/typography/layout), and apply the bundled Japanese slide-design principles digest during generation AND self-review. Trigger on スライド作成, スライドを作って, デッキ作成, プレゼン資料, 発表資料, slide deck, presentation slides, html2pptx生成.

Use when managing a YAML test-case inventory for a project: create test cases with GitHub Issues, track tests in test-board.yaml, decide what tests to run after changes, analyze regression impact, plan E2E tests, run Browser Use Cloud E2E, and test login/email authentication with Testmail-style disposable inboxes. Trigger on test-board, test-case-board, test-board.yaml, テストケース台帳, テストケース一覧, デグレ, 影響範囲, E2E, Browser Use Cloud, Testmail, ログイン認証テスト.

Search X (Twitter) posts, profiles, and threads via xAI's x_search API using a bundled Python CLI. Use when the user asks to search X/Twitter, check current discussion or reactions on X, find posts by specific handles, or research what is being said on X about a topic. Trigger on X検索, Xで検索, Twitter検索, x-search, X posts, ポスト検索, バズ調査, reactions on X.

Parent/global core skill for loop engineering shared by Codex, Claude Code, and child Skills. Use when a workflow needs autonomous or scheduled AI development loops, rigorous requirement clarification before execution, GitHub Issues, test-board.yaml, maker/checker separation, E2E evidence, durable state, or shared loop rules that multiple Skills should inherit. Trigger on 親スキル, loop-engineering-core, ループエンジニアリング共通, autonomous development core, requirements gate, scheduled agent loop.

Use when designing or running autonomous AI development loops that combine GitHub Issue-driven development, test-board.yaml test inventories, scheduled Codex/Claude runs, maker/checker separation, Draft PRs, E2E evidence, and regression impact analysis. Trigger on ループエンジニアリング, loop engineering, 自走開発, schedule/scheduled coding agent, AIが独立して実行, Issueとテストを合体, test-board plus GitHub Issues, unattended development loop.

Security-focused chaos engineering analysis that reads actual source code to identify vulnerabilities and failure-mode weaknesses. Use when the user asks to: (1) analyze a repository for security weaknesses, (2) run chaos engineering analysis, (3) find authentication/authorization flaws, (4) detect data leak risks, (5) identify failure scenarios like DB outages, session hijacking, or billing data loss, or (6) audit API rate limiting and cost exposure. Triggers on phrases like "chaos-engineer", "弱点を分析", "セキュリティ分析", "障害シナリオ", "脆弱性チェック".

Use when creating and publishing a Japanese note.com article for AI PLAY GUILD / PORSEO as a free public article with no membership plan, no paid separator, and no price. Uses the shared note article workflow for article writing, fact-checking, screenshots/diagrams, 1280:670 eyecatch generation, note draft creation, preview/editor verification, original publish, unpublished duplicate preview creation, and Discord delivery. Do not use for light-plan limited articles or Zoom-recording-sourced Pro articles; use note-light-publish or note-pro-publish instead.

Use when creating and publishing a Japanese note.com article for AI PLAY GUILD / PORSEO from a Zoom recording or Zoom-derived material as a Pro-plan limited article with one-off 2000 yen access. Uses the shared note article workflow for article writing, fact-checking, recording screenshots/diagrams, 1280:670 eyecatch generation, note draft creation, paid separator setup, preview/editor verification, original publish, unpublished duplicate preview creation, and Discord delivery. Do not use for free public articles or non-Zoom light-plan articles; use note-free-publish or note-light-publish instead.

Use when creating and publishing a Japanese note.com article for AI PLAY GUILD / PORSEO as a light-plan limited article with one-off 500 yen access. Uses the shared note article workflow for article writing, fact-checking, screenshots/diagrams, 1280:670 eyecatch generation, note draft creation, paid separator setup, preview/editor verification, original publish, unpublished duplicate preview creation, and Discord delivery. Do not use for free public articles or Zoom-recording-sourced Pro articles; use note-free-publish or note-pro-publish instead.

note.com の記事を企画・執筆・サムネ生成・投稿まで一気通貫で行う。ユーザーが「note記事を書いて」「noteに投稿して」と頼んだときに使う。

Substack の記事（長文ポスト）と Notes（短文ポスト）を企画→執筆→サムネ生成→公開まで一気通貫で行う。ユーザーが「Substack 記事を書いて」「Substack に投稿して」「Substack で発信したい」「Substack に Note 出して」「短文ポストして」と頼んだときに使う。

Substackの記事・短文（Notes）・チャットスレッドの操作をひとまとめにしたスキル。ユーザーが「Substackに記事書いて」「Substackに投稿」「Notesに流して」「Substackチャット返信して」「Substackの下書き見せて」など、Substackに対する任意の操作を頼んだときに使う。`substack-mcp` MCPサーバーが提供するツールを呼び出す。

Extract and summarize Claude Code conversation history from local transcript JSONL files, including chat-visible assistant/user messages, tool uses, tool results, session metadata, cwd, and git branch. Use when the user asks to recover, hand off, audit, or inspect what Claude Code did in a branch, worktree, repo, or previous session.

Use when Codex needs to configure, troubleshoot, or use X's official MCP servers: X API MCP at https://api.x.com/mcp through the xurl bridge, X Docs MCP at https://docs.x.com/mcp, OAuth2 setup, MCP client config for Grok/Cursor/Claude/VS Code, X API documentation lookup, or safe X API tool usage.

海外で成功しているWebアプリを発見→超詳細に分析→日本向けにローカライズ→ACP（Claude Code）セッションで実装するフルスタック自動化スキル。「日本向けにローカライズして」「海外アプリを日本で作りたい」「海外プロダクトを日本版にして」「ローカライズ開発」「Japan localize」「日本市場向けに」などと言われたら使う。

One-shot setup for an OpenClaw-style AI company knowledge workspace. Use when the user wants to bootstrap AI社員 / AIカンパニーナレッジ from scattered files, run the construction prompt, set up MEMORY.md and departments, or says「セットアップ」「まとめてセットアップ」「AI社員の土台を作りたい」「PCを調べてセットアップ」.

Generate slide decks (html2pptx, PPTX, presentation HTML) from a BUNDLED local archetype library — no external deck fetching. Pick a matching design archetype from the 18 bundled DESIGN.md systems by tag (mood × docType × industry × color), load its design tokens as the spec, generate slide HTML, then self-review with the Japanese slide-design principles before exporting via html2pptx. Use for product events, launches, pitches, IR, company intros, and any deck where you want a consistent, policy-safe design system without browsing real-world third-party decks. Trigger on スライド作成, スライドを作って, デッキ作成, プレゼン資料, 発表資料, イベント資料, slide deck, presentation slides, html2pptx生成.

Use for GitHub-centered AI development workflows: create or reuse Issues before coding, use issue-numbered branches, keep Japanese Issue and PR titles with 【repo名】 prefixes, write knowledge logs to Issues, manage PRs and GitHub Projects, and run local quality gates before completion.

Capture, preserve, and display evidence for E2E test runs, especially logged-in browser flows and browser test videos. Use when Codex is asked to run E2E tests, verify authenticated UI flows with video proof, use the computer-use:computer-use skill for logged-in screen inspection, leave evidence of an executed video test, show videos inline in the final response, collect Playwright/Cypress recordings, or produce a timestamped bundle containing test logs, videos, traces, screenshots, reports, and a manifest.

Search X (Twitter) through Hermes Agent's Python `tools.x_search_tool` directly instead of prompting the Hermes UI/agent loop. Use when the user asks to search X/Twitter, collect recent posts, inspect X discussion, retrieve post URLs, post text, like counts, authors, handles, dates, or source links, especially when they mention `x_search`, Hermes, Grok X Search, `uvx --from hermes-agent`, or want faster/rawer X search output.

Capture and reconstruct public webpages from a URL, including animation-heavy Framer, Motion, Lottie, RAF, scroll-driven, and video-driven pages. Use when Codex is asked to clone, reproduce, rebuild, archive, inspect, or high-fidelity recreate a landing page, marketing site, homepage, or other public web page by extracting live DOM, text, images, CSS, network assets, layout, responsive screenshots, generated JS motion hints, runtime animation traces, and browser-observable animations.