Productivity
git-coda
自动化 Git 工作流一键发布助手,执行完整的「ship」流程: 基于最新 main 切新分支 → commit → 推送 → 创建 PR → squash merge → 回 main。 调用 ship 本身就是对完整流程的授权;自动推断分支名、commit message 和 PR 内容, 并主动解决 Git 冲突、测试、lint、类型检查、构建和 CI 失败,持续修复和重试, 直到改动成功合并。不要因可自行修复的失败暂停,只在认证、权限、仓库保护等 无法由当前环境解决的外部阻塞出现时请求用户介入。 仅在用户明确表达「ship」意图时触发,例如: - 直接说「ship」「/ship」「git ship」 - 明确要求完整走完 git 流程:「帮我走一遍 git 流程」「拉 main、切分支、提 PR、合并」 - 使用类似「发布这次改动」「把这些改动提成 PR 并合并」的表达 不应触发的情况: - 只是想 commit - 只是想创建 PR - 只是想 push 代码 - 讨论 git 相关问题或解释 git 概念
$ npx skills add rufus-fan/git-codaquestion-drafting
Draft, audit, or expand the buyer questions to track in AI apps (AEO or GEO prompt tracking): branded and unbranded questions, intent clusters, and real buyer phrasing. Use for "which questions should I track in ChatGPT", a review of a question set, or new questions for a Frontmost AI project. Adds questions only after an explicit go-ahead.
$ npx skills add frontmost-ai/aeo-skillssemantic-html
Reviews, rewrites and explains HTML5 semantics (tags, landmarks, headings, forms, ARIA). Use whenever the user pastes HTML or asks for a markup review, the right tags, accessibility or fixing 'divitis'.
$ npx skills add lucasaugustsof/skillsprd-ux-logic
Write UX logic acceptance criteria in Gherkin format for any product initiative. Works with or without a prd-design handoff. Produces Given/When/Then tables grouped by flow, with naming conventions defined first.
$ npx skills add elleong58-hash/design-skillspic-firmware
Microchip PIC and dsPIC firmware with MPLAB X IDE, XC8 / XC16 / XC-DSC / XC32 compilers and MPLAB Code Configurator (MCC Melody/Classic) for PIC16, PIC18, PIC24, dsPIC33 (F/EP/EV/CK/CH dual-core/A) and PIC32 — configuration bits, oscillator/PLL, TRIS/LAT/PORT/ANSEL, PPS, interrupts per family, 8/16-bit atomicity and hardware-stack limits, dsPIC DSP engine and Q15 fixed-point, high-speed PWM/ADC control loops, motor control and digital power, XC8 memory and optimization, headless builds, programming and debugging with PICkit 4/5, MPLAB Snap and ICD (ipecmd), reset causes and traps. Use this skill whenever the user mentions any PIC or dsPIC part (PIC16F, PIC18F, PIC24F, dsPIC33EP/CK/CH, PIC32MX/MZ), MPLAB X, MCC, XC8/XC16/XC-DSC/XC32, PICkit, Snap, ICD, `#pragma config`, motorBench, or a .X project folder — even for a short question like "o pino não lê nada no PIC" or "o meu controlo PI no dsPIC oscila". Use together with embedded-c-cpp for general C quality, testing and arc...
$ npx skills add budenga/firmware-skillsacademic-visual-slides
Create or refine editable academic and teaching presentations from scripts, annotated PDFs, and visual references. Use for clear university-style slides with mint, blue and muted orange regions, transparent cartoon illustrations, concrete examples, and distinct opening, takeaway and closing pages. Not a rule to redesign every presentation or replace a required institutional template.
$ npx skills add LandyLee-gdut/academic-visual-slidesc1-comms
Rework or critique something the user has already drafted (email, Slack message, weekly or status update, decision memo, proposal, deck outline, PR description, incident summary, design doc) so a busy senior reader can act on it, with the main point first, an explicit ask, facts kept apart from judgment, risks in plain sight and nothing invented. Use whenever the user asks for the "C1 lens", "C1 comms" or "Capital One style", or wants a draft of theirs made exec-ready, answer-first or BLUF, reworked for a VP, director, leadership, board or steering-committee reader, or pressure-tested before it goes upward ("what will they poke holes in?"), even if they never mention Capital One. Takes pasted text or a file path. Not for text you are writing yourself in the course of other work, and not for writing something new when there is no draft or notes to start from.
$ npx skills add commahawk/c1-commsarticle-review-panel
Reviews a technical article, blog or LinkedIn post through three independent expert lenses (domain expert, editor, statistician) and writes one markdown file per lens flagging what is good, what is wrong and what needs more explanation, verifying every number against the raw data next to the draft. Use when the user says "review my post", "act like an experienced JVM engineer and review this", "review this article as an editor", "check the statistics in this post", or "is this post ready to publish". Do NOT use to review code or a pull request, use diegourban:gh-pr-review, diegourban:branch-review or /code-review instead. Do NOT use to fact-check a claim about how a codebase works, use the diegourban:fact-checker skill instead. Do NOT use to write the post from scratch, have a direct conversation instead.
$ npx skills add diegourban/claude-skillsRobTF9/meeting-notes
Turn a meeting transcript into short notes with decisions, actions and open questions.
$ npx skills add RobTF9/skill-grouping-demo-2RobTF9/onboarding-checklist
Make a first-week checklist for a new teammate.
$ npx skills add RobTF9/skill-grouping-demo-2Brajesh31/claude-daily-meeting-update-skill
Interactive daily standup/meeting update generator. Use when user says 'daily', 'standup', 'scrum update', 'status update', 'what did I do yesterday', 'prepare for meeting', 'morning update', or 'team sync'. Pulls activity from GitHub, Jira, and Claude Code session history. Conducts 4-question interview (yesterday, today, blockers, discussion topics) and generates formatted Markdown update.
$ npx skills add Brajesh31/claude-daily-meeting-update-skill20vision/skills
Agent skills: WCAG audit, house plans, floor plan CAD, cloud computer tasks
$ npx skills add 20vision/skillssekretmaksym-dev/brief
Writes BRIEF.md before a website is designed, in one of two modes the user picks. Site mode gathers what the page needs (offer, audience, proof, sections, style inputs) and hands off to /style-guide without business advice. Diagnose mode starts from the user's real situation (nothing yet, a niche, a freelance account or service with few clients, a live site), researches demand and competitors, stress-tests the idea with independent reviewers, and may conclude that a site is not the next step. Use when the user invokes /brief, wants to start a website or landing page, asks what site or business to build, wants a site that makes money but has no idea, wants a niche or idea validated, asks whether an idea is worth it, or asks why they get no clients. Not for generating tokens or building pages.
$ npx skills add sekretmaksym-dev/design-skillsanki1kr/idea-truth
The one skill for "will this succeed, and why not?", for ideas AND live products. Mode IDEA (before deck, plan or code; hackathon, startup, feature, research topic) and Mode PRODUCT (an existing site/app). Use when the user says "idea check", "is this worth building", "research my idea", "roast", "roast my landing page", "why no signups", or pastes a problem statement, names an idea, or gives a product URL. Runs like experienced founders, PMs, investors and judges: kill-questions, a skeptic round, real-behaviour evidence, incumbent teardown, economics, a 41-check failure-mode sweep, pre-mortem, then a verdict (DEAD/UNKNOWN/ALIVE) or a graded product roast with an ordered fix list. Every source opened, every claim labelled.
$ npx skills add anki1kr/second-natureCreativityHouse2025/project-rescue
Diagnoses a project that is in trouble and builds an honest recovery plan - separates symptoms from root causes using the project's own documents, sizes the gap in time, cost and scope, lays out the real options (recover, reduce scope, add time or money, pause and re-plan, stop), and writes a one-page brief for the sponsor, in English or Arabic. Use this whenever someone says their project is late, over budget, stuck, failing, "red", out of control, or losing the client's or sponsor's confidence; has inherited a troubled project; needs to tell leadership bad news about a project; or asks how to get a project back on track, even if they do not use the word "recovery". Not for planning a healthy project from scratch, routine status reporting, or resolving personal conflicts between individuals.
$ npx skills add CreativityHouse2025/pm-skills-for-claudeAigine/aigine-redact
Mask personal data (GDPR) and confidential information (corporate trade secrets, procurement tenders, FOIA exemptions, and statutory secrecy such as Swedish OSL) in PDF, Word, PowerPoint or image files using specialized non-generative encoder AI in the EU (98.30% precision, 93.17% recall, 95.67% F1; interactive web review GUI; strict zero retention) via https://redact.aigine.se/mcp. Use when a document that may contain information about identifiable people, trade secrets, or public/statutory secrecy is about to be sent to an external AI model or service, shared, or released, or when the user asks to redact, anonymize, pseudonymize or mask a document.
$ npx skills add Aigine/aigine-redactspace-enthusiast/meeting-minutes
Turn a meeting recording (m4a, mp3, wav, mp4, ...) into minutes on the user's own computer. Whisper transcribes, pyannote separates speakers, and the agent writes minutes from the speaker-labelled transcript. Use when the user gives a recording and asks for minutes, a summary, a transcript, speech-to-text, or "who said what". 회의 녹음으로 회의록, 전사본, 화자 분리, STT 를 요청할 때 사용한다. The audio never leaves the computer.
$ npx skills add space-enthusiast/meeting-minutes-skillhaslanli/glance
Use this skill whenever the user responds to an answer Claude already gave (or to several answers, or to pasted text) by asking for less of it rather than more: the bottom line, the gist, the conclusion, tl;dr, the key points, a shorter version, a simpler explanation, or a visual of it. Also use it when they say the answer is too long or confusing. Requests in any language count (\"покажи кратко\", \"长话短说\", \"要点だけ\", \"一言で\"). Also use it for a recap of everything done or decided so far in the session. The skill re-tells the same content: the answer in plain words first, then a chart, tiles or table of its data, then short bullets. It keeps every number, name, command and warning, invents nothing and cuts wording, not facts: text ends up about half to two-thirds as long; tables keep their size. Commands: /glance, /btw /glance, /glance [what|on|super|off|status|help]. Don't use it for new deliverables such as documents, dashboards, slides, emails, exports or translations, or for new questions about the topic.
$ npx skills add haslanli/glance-skillXs-trek/frame-align
在动手实现之前,通过可交互的 HTML 页面与用户对齐一个软件想法或难题的框架:模型用自己的话复述理解、提供灵感、指出空缺与矛盾、定位难点,用户用点选和备注回应,直到用户认为框架成立。仅在用户明确调用(例如 /frame-align)时使用,不要自行触发。
$ npx skills add Xs-trek/frame-align-skillkeyz182/adversarial-review
This skill should be used when the user asks to "challenge this plan", "red-team this", "what could go wrong", "pre-mortem this", "adversarial review", "poke holes", "stress-test this design", "find the flaws", or "review this before I build it". It looks for the reasons a plan, design, diff, decision or another skill's output will fail, ranks the concerns against a fixed severity bar, and ends with one verdict (PROCEED, AMEND or RETHINK). Runs interactively by default, or in report mode with no questions when called by a worker agent or when the caller asks for "report mode".
$ npx skills add keyz182/keyz-claude-skillsSharkByte561/windows-troubleshoot
Diagnose and troubleshoot Windows 10/11 and Windows Server problems with Microsoft's TroubleShootingScript toolset (TSS) - map a symptom to the right TSS scenario and exact .\TSS.ps1 command, run a read-only health triage of the local machine (event logs, crashes, BSODs, disks, services, updates, network, domain), and analyze an already-collected TSS / MS_DATA zip into plain-English findings. Use this whenever the user describes a Windows problem (slow PC, crashes, blue screen, account lockouts, logon or GPO problems, DNS/Wi-Fi/VPN/SMB/proxy connectivity, printing, RDP/AVD/Windows 365, Windows Update or activation failures, Start menu/Store/app issues, WMI/WinRM, Hyper-V, clustering), asks "which TSS scenario", "what does TSS collect", "how do I trace X", mentions TSS, MS_DATA, xray or SDP, asks whether TSS is expired or wants it updated, or hands over a support-data zip or .evtx files to read - even if they don't name TSS.
$ npx skills add SharkByte561/windows-troubleshootBelweave/belay
Reach any goal a command can score (fix every failing test, finish an audit, make something faster or smaller, port a tool, polish a UI or a text) by climbing in small verified steps. Each change is scored by a locked check, committed if it wins and undone if it doesn't, across one or many repos. Use when the user says "belay", "hill climb", "climb this", or "keep going until it passes", wants a long unattended run toward a measurable target, or asks to resume or report on a climb.
$ npx skills add Belweave/belaypriscilarwagner/meeting-digest
Digest all meetings from a given day — uses the calendar as source of truth and the configured meeting-notes sources as content. Extracts decisions, action items, and blockers, writes structured notes to Context/Meeting Notes/.
$ npx skills add priscilarwagner/pmshippingbasamakcomtr/support-assistant
Triage customer support requests, identify missing evidence, suggest safe next actions, and draft CRM tasks and customer replies. Use when assessing support messages, preparing responses, or turning requests into actionable internal tasks. Never imply that a message was sent or a ticket was created unless it actually happened.
$ npx skills add basamakcomtr/basamak-skills