Edição de reels no estilo @agreycealves com GPU NVIDIA (Linux): recorte e Whisper na GPU + upscale com IA (Real-ESRGAN) para vídeos do WhatsApp.
1 skill
Claude Code and Codex plugins their authors filed under "productivity". All categories.
Edição de reels no estilo @agreycealves com GPU NVIDIA (Linux): recorte e Whisper na GPU + upscale com IA (Real-ESRGAN) para vídeos do WhatsApp.
1 skill
Claude plan usage (5-hour and weekly, % used, time to reset) in the Claude Code status line and the macOS menu bar.
hooks
Claude plan usage (5-hour and weekly, % used, time to reset) in the Claude Code status line and the Windows tray.
hooks
Read-only researcher agent that answers one concrete question from the codebase or external sources, with cited evidence
1 agent
Writes, reviews and debugs CasparCG AMCP commands and the Bitfocus Companion buttons that send them (lower thirds, stingers, intros, transitions, mixer, templates), with facts tested on CasparCG 2.5.1.
1 skill
Generates a set of HTML lower thirds and event graphics (speaker, timer, ticker, announcement...) for CasparCG from client materials, driven by JSON and live Companion / Ontime variables.
1 skill
Generate natively editable PPTX from PDF / DOCX / URL / Markdown — real DrawingML shapes, text boxes, charts, and animations. Note: run `pip install -r requirements.txt` inside the installed plugin directory for the Python post-processing scripts to work.
1 skill
Audit a DES, system dynamics, agent-based, or hybrid simulation write-up against the STRESS reporting checklist (Monks et al. 2019), item by item with quoted evidence.
1 skill
Delegate bulky, low-judgement work to opencode's best free model so it costs zero primary tokens. Dispatches the task, returns a compact digest, and tracks exactly which files the worker touched.
1 skill
Ends a turn left on an open dialog (permission prompt, question) for 30 min by default and tells the agent the user is /afk; /dialog-timeout <minutes>|off|on
1 skill
Nudges the agent on when it stops early, and derives DECISIONS.md from your own sessions; /stop-judge
hooks
Coordination across the agent fleet: the AI scrum, a round table of what every session, subagent and pull request is doing, with the collisions, stale lanes and decisions waiting on the user
1 skill
Apply the human-centered language standard to any writing that describes people. Use when drafting or revising appeals, newsletters, social posts, testimony, case notes, press releases, or internal memos for an organization that serves, represents, or advocates for people. Applies person-first language, removes deficit framing, keeps the organization out of the hero role, and defaults to plain language. Also use when asked to rewrite respectfully, check for stigmatizing language, or make copy person-first.
1 skill
Apply the trauma-informed standard (SAMHSA framework, editor Laura Braden) on top of the human-centered language standard. Use for any writing by or for an organization serving people affected by trauma: shelters, treatment providers, family services, survivor advocacy, community health, and their funders. Covers fundraising appeals, donor stories, newsletters, social posts, case notes, summaries, and crisis communications. Use when asked for trauma-informed writing, when donor storytelling involves people served, or when a situation is urgent or distressed.
1 skill
Suppress common AI writing patterns so drafts read as written by a person. Optional add-on to the human-centered language standard. Use when polishing appeals, newsletters, social posts, reports, or any copy that should not sound machine-generated. Also use when asked to make copy sound human, remove AI tells, or de-AI a draft. Does not replace human-centered or trauma-informed practices if those are also loaded.
1 skill
查询 AI HOT 中文 AI 资讯、精选、当前热点和日报。通过 aihot.virxact.com 公开只读 API 获取实时数据,无需 API Key。
1 skill
横纵分析法深度研究:纵轴追溯生命历程,横轴同期竞品对比,交叉产出洞察,最终生成排版精美的 PDF 研究报告。
1 skill
数字生命卡兹克公众号长文写作风格:根据素材撰写、续写、扩写公众号长文,保持活人感与真诚表达。
1 skill
ユーザのメモ書きや箇条書きから、ADR(Architectural Decision Records)の新規レコードをMarkdownファイルとして作成するスキル
1 skill
Word・PDF・PowerPoint・Excel・Markdown など任意形式のドキュメントを、誤字脱字から目的適合性・全体構成まで多層的にレビューし、優先度順の推奨事項をまとめた Markdown レポートを作成するスキル
1 skill
Persistent local memory shared by all your AI assistants: what you told them, kept as dated facts in one SQLite file on your machine and recalled before they answer. Runs locally from a bundled binary for macOS Apple silicon and Windows x64; no account or service.
1 skill
ユーザの背景情報とやりたいことのざっくりとした概要を入力として受け取り、ヒアリングを通じて本当の課題を特定したうえで、効果的な解決策を提案する壁打ちスキル
1 skill
Read, search, triage and draft email over IMAP and SMTP. Runs locally; sending and permanent deletion are off by default. The Claude Code plugin uses a native bundle for macOS Apple silicon and Windows x64.
1 skill
CLI + skill: operate your Grok Bot desktop app from Claude Code on your Mac: agents, groups, sections, pins, transcripts, send --wait. Unofficial; needs Node 20+.
1 skill
Skill: for a Grok Bot agent inside its own cloud box: manage sibling Grok Bot workers (list, create, organize, send, delete) under an advisory agent policy. Unofficial.
1 skill
Skill + helper: operate the Claude Desktop app on your Mac from Claude Code: tasks, replies, approvals, projects, skills, connectors. Unofficial; needs macOS 15+ and Peekaboo.
1 skill
Academic study work by data: anti-AI writing checks, APA citations, concept maps, typeset PDFs and decks, free-image sourcing, multi-area workspaces.
1 MCP server
Anthropic: PR-Review mit mehreren Agents und Konfidenz-Filter; postet das Ergebnis per gh als Kommentar in den Pull Request.
1 command
Word and punctuation choice. Removes the vocabulary and punctuation patterns that mark writing as AI generated.
1 skill
What a message says and how it is shaped: length, ordering, prose versus lists, and whether a sentence carries information.
1 skill