mkbhd-voice
Conversational video-script voice in the style of a tech YouTuber explaining to a smart friend. Use when the user asks for mkbhd style, conversation style,…
Skills
Find an Agent Skill for your task. Read the guide and copy the install command.
Updated 11-02 00:09
Writing & research
14,601 results
gtb
Search and browse books from Project Gutenberg directly in the terminal. Core Scenario: When the user needs to find classic literature, read full book text, or…
wkp
Wikipedia search and summary extraction tool. Core Scenario: When AI needs to quickly obtain definitions or official summaries in the terminal.
writer
AI text processing module for simplified translation, summarization, and refinement. Core Scenario: When the user needs to translate, polish, or summarize text…
documentation-templates
Documentation templates and structure guidelines. README, API docs, code comments, and AI-friendly documentation.
flutter-documentation
Generate and maintain project documentation for Flutter apps: README, architecture docs, feature docs, and onboarding materials. Use when the user wants to…
readable-japanese
A skill for tuning Japanese sentences into a rhythm that is easy for humans to read. Based on the mechanism by which the brain processes sentences by "putting them to voice" during silent reading (inner prosody / subvocalization), it checks and fixes phonological confusion, prosody-syntax alignment, and sentence monotony. When writing, polishing, or reviewing Japanese sentences, or for "hard to read," "bad rhythm," "stiff," "how to place commas"...
article-illustration-planner
Deeply analyze the article's structure and precisely plan illustration placement and visual imagery. For knowledge/tutorial articles, the "Cognitive Load Reduction and High-Fidelity Diagram Replacement for Text" mechanism is enabled by default; deleting formulas and quantitative metrics is strictly prohibited; 165 layout diagram types (IG-* infographics / SC-* social media cards) and 331 hand-drawn styles/36...
specify
Turn a goal into an implementation plan (spec.md). Simplified layer chain: L0:Goal → L1:Context → L2:Decisions → L3:Requirements → L4:Tasks. Evidence-based…
writer-memory
Agentic memory system for writers - track characters, relationships, scenes, and themes
ljg-paper-river
Paper reverse-reading method: given a paper, recursively find the prior papers it critiques and improves upon (up to 5 levels), then find the latest progress after it, and narrate the history of the problem's evolution forward from the source. With the problem as the axis, explain in Feynman style the problem each paper saw and its innovation in solution. Use when user shares a paper and wants to understand its intellectual…
zhiforge
ZhiForge - Intelligently transforms the knowledge base (Markdown blog) into Zhihu answers and automatically publishes them. Supports searching trending questions, matching the knowledge base, writing articles, automatic publishing, quality review, and saving back to the knowledge base—a complete closed loop.
convert-to-natural-writing
Use skill if you are humanizing or rewriting AI-sounding, robotic, or generic text, Markdown, MDX, or HTML into natural multilingual copy; not…
research-powerpack
Use skill if a coding task needs fresh external research from Google, Reddit, scraped docs, or deep structured synthesis.
run-comprehensive-research
Use skill if you are orchestrating multi-domain research across official docs, community forums, and practitioner experience with parallel researcher agents…
youmind-article-dispatch
Dispatch content to multiple platforms from a single topic — Dev.to, Hashnode, WordPress, Tumblr, Ghost, Beehiiv, Kit, X/Twitter, WeChat, Qiita. Each platform…
grill-with-docs
Relentless interview that sharpens a plan and updates CONTEXT.md and ADRs as decisions land.
modeling-paper-rubric-and-model-selector
Generate a modeling paper structure and writing checklist based on common scoring points, and provide model selection and comparative experiment routes according to the problem type and data conditions. Invoke when there is a need for "paper format/scoring alignment/model selection/route uncertainty".
paper-micro-unit-generator
Batch-generate and merge paper content based on micro-unit templates and scripts. Invoke when it is necessary to split writing into micro-units, write unit by unit, and automatically generate the paper with scripts/generate_all_offline.py and scripts/merge.py.
paper-workflow-orchestrator
MathModel Skill main entry point. Trigger words: mathematical modeling, generate paper, analyze competition problem, CUMCM, MathorCup, Huashu Cup, MCM/ICM, ICM, MathModel. For any mathematical modeling task, read this skill first, then route to sub-skills.
problem-doc-model-selector
Parse the competition problem PDF/Word, extract tasks and data conditions, and provide model selection and validation route. Invoke when the user provides a competition problem document or problem statement text and needs to determine which model/method to adopt.
novel-game-analyze
Deconstruct a novel for game adaptation. Compress a raw novel, deconstruction library, or writing project into a SOURCE_BIBLE with cited textual evidence,…
short-drama-video-prompts
Write the short drama storyboard and frozen keyframes as directly copyable video prompt Markdown, and can also write timeline score/theme song intent as requested by the user. Use when the user mentions text-to-video/image-to-video actions, character performance, camera movement, lip sync, environmental motion, shot duration, start and end states, converting storyboards into video prompts, or writing score prompts; do not generate media, do not create lyrics, do not change storyboard boundaries.
short-drama-write
Write or revise shootable Chinese short-drama and comic-drama single-episode Markdown scripts, and also normalize existing scripts while preserving the author's original text. Use when the user requests "write/revise one episode of a short drama," "turn an outline into a script," "optimize scenes/dialogue," "remove template feel," "polish away AI feel," "continue to the next episode," or provides script requirements to enter subsequent production; not responsible for assets, storyboards, media prompts, or final review.