Productivity

Import GTFS transit feeds, query station/stop/pathway/route/trip/calendar/shape data, edit connections, nodes, routes, trips, and stop times, compare trip service patterns and schedules, export changes to GTFS CSV, and open a local browser dashboard with the gtfs-tools CLI. Use when working with GTFS data, transit stations, pathways, routes, trips, schedules, or accessibility audits.

Session-opening routine for a project with history, or for an umbrella project that holds several: gathers what earlier sessions left behind (a project manager such as Notion, a session log if the setup writes one, git, an Obsidian vault, handoffs, chats, memory and project files), cross-checks it, summarises where everything stands, asks what to work on today and hands the answer over to a skill-selection skill if one is installed. Chooses its sources from the tools available in the session, not from the name of the environment. Use it when opening a session even if the skill is not named, with phrases such as "let's pick up", "where were we?", "carry on with my-app", "open a session on X", "let's start", "/resume-project", or the equivalent in the user's language (in Spanish, for example, "retomamos" or "¿por dónde íbamos?"). Do not use it midway through a session that is already working, or to create a project from scratch.

Work-planning routines for people who run several projects in parallel: a daily plan (today's blocks, each with its next action and a first step), a day close (where each block was left and the next actions for the next working day) and a weekly review (what went well, collisions, priorities). Reads the schedule, rules and project status from the user's project manager and project instructions, in claude.ai or in Claude Code. Use it whenever the user says 'what's on today', 'what should I do today', 'plan my day', 'where do I start today', 'I'm done for today', 'closing the day', 'weekly review', 'what have I got this week', or the equivalent in their language (in Spanish, for example, 'qué toca hoy', 'cierro el día' or 'revisión semanal'), even if the skill is not named. Do not use it to open a session on a project with history (that is a session-opening skill such as resume-project), to archive or close the chat, or to choose skills for a task.

Free up disk space on macOS by safely cleaning regenerable caches, old runtime versions, and bulky one-off artifacts. Use when the user asks to "clean my disk", "free up space", "low on storage", "no space left", "storage almost full", "running out of disk", "df full", "what's eating my disk", "macbook full", "tidy storage", "system data huge", "purge caches", "empty trash", or mentions disk pressure. Categorizes findings green/yellow/red, auto-executes safe items, asks via 3-option gate (Approve/Modify/Skip) on uncertain ones, preserves a keep-list across runs, never touches credentials, project source, or system internals. Single-machine state in skill dir; resumable mid-run.

/checklist installa nella cartella corrente ciò che manca per la checklist, poi la mostra in una chat dedicata. Non sovrascrive e non cambia nulla di ciò che c'è già.

Build a study flashcard deck from any subject, or from material the user already has (lecture PDFs, slides, Word or Markdown notes, a course folder, a repo, or existing Anki or Quizlet cards), by first asking the user what they need. Produces print-friendly A4 landscape question-and-answer card images grouped by topic, one PDF per topic, and a README index, and on request YouTube videos (question then answer, a questions-only quiz, or a cards-only review) with chapter descriptions and thumbnails. Use this whenever the user asks for flashcards, study cards, review cards, revision cards, a Q&A deck, quiz cards, or "cards to remember" something, or wants to turn a codebase, course, chapter, or topic into cards, printable cards, or a flashcard video, even if they do not say "deck". Also use it to add, fix, split, or regroup cards in a deck made with this skill.

Run a large engineering project with agents as a campaign. The lead (the main session) splits the owner's goals into quests, each one independently verifiable outcome; contributor agents complete them on their own branches and worktrees; a fresh reviewer examines each quest as its risk requires; and the `campaign` tool keeps the state, runs builds, tests and benchmarks as jobs on exact commits under `quiet`, and lands only what was verified. Use when starting or resuming a multi-part project (an engine, a library family, a migration, a rewrite), when work should run in parallel worktrees, when a quest needs review or is ready to land, or when the owner says "start the campaign", "resume the campaign", "status", "review it" or "land it".

Publish a local wpsync copy of a WordPress site to the live site, replacing it; roll back or finalize afterwards. Use only when the developer asks to publish, push, deploy, go live, roll back or finalize.

Local development of live WordPress sites with wpsync and DDEV, from only an admin login. Use when the developer says "new wp-dev project", "I want to work on my WordPress site locally", "set up example.com", "continue my wp-dev project" or "continue wp-dev", or asks where a wp-dev project stands. Works out the phase (set up, work, publish) and runs it.

Work with the user in English to shape product requirements, build reviewable prototypes from agreed flows, and turn feedback into specific, verifiable changes. Use for product discovery, requirements discussions, prototype reviews, and product revisions. Handle simple copy edits or small, well-defined code fixes directly.

Convert scanned documents, handwritten notes, images, and PDFs into readable typed, searchable PDFs with simple formatting and approximate layout. Use for everyday digitization when exact line breaks, spacing, columns, and annotation positions are not required.

Turn work into a structured, interactive HTML report and keep it in step with the session: a work map of tasks, and for every task — and every stage inside it, recursively — the problem, the goal, the context gathered (with links), conclusions, assumptions, what was not checked, the result, a workflow canvas of stages, what needs the reader's decision and what the reader must do. The reader can answer the agent's questions inside the report and send them back to continue the session; a prompts r

Turn trip materials (flight/hotel screenshots, conference agenda, tickets, links) into a mobile-friendly travel cheat-sheet web page with an itinerary timeline, checklists, destination handbook, dual-timezone clocks, map deep links, QR ticket pages and an optional password gate; outputs a single-file HTML page or a multi-page static site ready for GitHub Pages. Use when the user asks for a travel itinerary, trip planner, conference trip preparation, business-trip plan, 行程规划, 出差/参会行程, 行程攻略, 随身页,

Take stock of the whole harness in one report and one approval, then apply on a branch (ハーネスの棚卸し).

Audit the agent definitions in a Claude Code harness (~/.claude/agents or a repo's .claude/agents) and give each a verdict.

Core rules for the todo-* skills (todo-init, todo-plan, todo-add, todo-continue, todo-update, todo-status, todo-report, todo-release) - the task tree in TODO.md, compact task format, complexity levels and model routing, IDs, git flow with a no-git fallback, invariants, Definition of Done and hard guardrails. Load this before running any todo-* command or when editing a TODO.md in the `# Active Backlog` task-tree format. Invoked directly as /todo it only prints the command list and a...

Operate SiYuan Note CLI (siyuan) to manage workspaces, notebooks, documents, blocks, SQL queries, search, export/import, assets, attributes, bookmarks, tags, history, refs, repo snapshots, sync, serve, database attribute views, templates, outlines, and system info. Use this skill when the user mentions SiYuan, 思源笔记, or asks to query/manipulate SiYuan data via CLI.

Coordinate a large piece of work across several parallel Claude Code desktop sessions, with the current session acting as the coordinator. First interviews the user until the requirements are fully clear, then splits the work into independent tasks, launches each one as its own session (own git worktree and branch), supervises their progress, launches dependent waves, verifies results, and shows a live visual dashboa…

协调跨阶段开发、恢复阶段尚不明确的任务，或记录与复核任务过程中的决定和证据。用于从想法到上线、多阶段迁移、自主执行和决策日志审计；明确规格的实现使用 dev-build，追查某项设计当时的理由使用 dev-why。

A template skill for preparing your own German income tax return (Einkommensteuererklärung, tax years 2024 onward). It starts with a 12-question interview that finds which areas apply and what to gather, then runs a gated workflow where scripts, not the model, match claimed amounts to their invoices and apply common legal limits.

会議の名前から、その打ち合わせの今日までの経緯・課題・宿題と今後の予定を、カレンダーと資料から要約する。

A HIPAA-supporting initial-consultation booking system for a SOLO therapy private practice, built on Google Workspace + Apps Script + Sheets + Calendar. Use when a solo practice owner wants (1) a public booking page where prospective clients book a free initial consultation against the owner's real calendar availability, with confirmation / reminder / follow-up emails and a private reschedule-cancel link; and (2) one private owner console to set weekly consult hours, date-specific hours and time off, view upcoming consults and mark no-shows, and edit email templates and booking rules. Adapted from a group-practice system (public Booking app + Manage Availability + My Consults + Consult Admin) — the solo edition merges the three staff apps into one console and drops the group edition's calendar-connection app entirely.

Coordinates a collaborative multi-agent team for large software engineering projects, multi-file refactoring, systems simulations, document reviews, and mathematical proofs. Conducts a Phase 1 scoping interview (defining objectives and verifiable invariants with objective verification and integrity modes) and executes Phase 2 autonomous multi-agent orchestration with Sentinel, Project Orchestrator, Explorer, Worker, Critic, Challenger, Auditor, and Victory Auditor roles. Use when the user invokes /teamwork or requests a team of autonomous agents to deliver a complex project.

Plan and explain systems through a saved, clickable map. Use for collaborative design, continuing a planning conversation, system explanations, or preparing an agreed plan for implementation. Keep design questions in chat.