В каталоге 182 820 навыков в 14 категориях.
Apply R&D framework to optimize prompts and context. Use when optimizing context window usage, reducing prompt size, del...
Auto-edit a raw talking-head clip into a finished vertical 9:16 reel — local transcription, word-synced burned captions,...
Benchmark LLM reference recommendation capabilities by verifying every cited paper against Crossref, PubMed, arXiv, and...
TOON format knowledge and usage patterns for agent communication and memory persistence in plan-marshall marketplace
Create and organize reference files for complex skills with 40+ patterns. Guides sub-topic identification, file structur...
Iteratively refine image generation prompts based on feedback about previous renders. Use when the user wants to improve...
This skill should be used when the user asks to "refine a prompt", "optimize a prompt", "improve my prompt", "rewrite pr...
Iteratively review and revise a prompt plan until no new findings survive evaluation. Runs /review-prompt-plan, /evaluat...
Refine vague or unclear prompts into precise, actionable instructions. Use when user asks to clarify or improve instruct...
Move a skill from staging to active
Use when the user says 'we decided', 'I realized', or 'that's a pattern' — captures a share-ready decision, finding, or...
Silently refresh AI context by reading project configuration and guidelines. Use when starting a new conversation, after...
Review recent changes and update CLAUDE.md to keep it relevant for future sessions. Use after significant changes or per...
Update stale workspace context.
Reads and loads architectural context files (architecture.md and architecture.puml) to understand the PSX Memory Card Ma...
Decision framework for choosing between regex and LLM when parsing structured text — start with regex, add LLM only for...
Maintain the skills registry by scanning installed skills and updating skills.json. Use when adding, removing, or auditi...
Regulatory Drafting agent for healthcare workflows.
Run a bounded REINFORCE-style policy update with terminal RLAIF rewards, a value baseline, and KL regularization evidenc...
Compact the current Codex VS Code extension thread through Rekall with a verified handoff and optional one-time continua...
Find all notes related to a topic using parallel sub-agents for comprehensive search.
Discover and install related skills from inference.sh skill registry. Helps find complementary skills for your AI workfl...
Hand the session to a fresh one instead of compacting it: write a summary of the conversation to a committed file, start...
Capture typed feedback about skill mistakes so skills can learn and improve over time. Writes structured entries to `~/....
Memory consolidation and defragmentation for long-term memory maintenance. Use when asked to consolidate memories, defra...
Summarise current status, completed work, and remaining tasks when context is running low. Use when you need to hand off...
Transparent-PNG cutouts / background removal for flat illustration, sticker, cartoon and avatar art — rembg (isnet-anime...
Transforms lessons learned into domain-organized memory instructions (global or workspace). Syntax: `/remember [>domain...
Store learnings in memory system for future sessions
Quickly save a memory to the soul
Save the current work session to persistent memory for future context. Summarizes accomplishments, tracks files modified...
This skill should be used when the user says "remember", "note this", "add to CLAUDE.md", "don't forget", "keep in mind"...
Capture friction as methodology notes. Three modes — explicit description, contextual (review recent corrections), sessi...
Store a learning, pattern, or decision in the memory system for future recall
Reloads critical instructions when behavior degrades. Use when forgetting tools, not delegating, or missing safety rules...
Store a learning, pattern, or decision in the memory system for future recall.
Advanced memory operations reference. Basic patterns (profile loading, simple recall/remember) are in project instructio...
ALWAYS USE THIS SKILL WHEN STARTING ANY KIND OF WORK, NO MATTER HOW TRIVIAL. You have no memory between sessions and wi...
Use when launching, controlling, documenting, debugging, or cleaning up remote Linux GUI apps through a clean Xvfb deskt...
Capture professional-quality remote interviews using double-ender technique and dedicated recording platforms for podcas...
Programmatic video creation with React via Remotion — compositions, animations, motion graphics, rendered to MP4. Includ...
Writes detailed video scripts for Remotion based on input requirements and codebase analysis
> 目标：安全地移除一个 SKILL（推荐废弃而非直接删除）
Remove backgrounds from images — background removal API for transparent PNGs, cutouts, and masks. Segment foreground fro...
Remove one named unwanted object from a supplied image and reconstruct the revealed scene with an AI-native image edit w...
Identify and remove phrasing, structure, and word choices that signal AI authorship. Runs iterative passes through fresh...
任正非视角. 华为创始人 / 总裁, 东方 CEO OS 流派 (流派 E) 奠基样本. 从心声社区讲话 / 《华为的冬天》/ 《以灰度看世界》/ 公开长访谈蒸馏的 founder-led 公司经营镜片: 灰度哲学 (方向大致正确 + 组织...
Assemble a viral iOS "AirDrop" notification-carousel video ad (≈6–8s, 9:16) from a brand line plus 6–16 real product pho...
Assemble a cinematic live-action-style music-video ad from a config — an original sung anthem carries the whole narrativ...
Assemble a creator picture-in-picture product-listicle ad from a config — the creator stays FULL-FRAME the whole beat (v...
Assemble an iMessage notification-cascade video ad (≈14s, 9:16) from a phone-on-desk plate + 3–5 messages — authentic Ap...
Assemble a silent, music-led 3-world product-tour ad — trim and hard-cut-concat the per-world WIDE-arrival + top-down-ma...
Render a punchy ~12s vertical (9:16) music-only direct-response OFFER ad as a 4-beat kinetic-typography film — HEADLINE...
Build the deterministic PIL pill overlays for a 'perfect-score + proof-points' UGC video ad (white 10/10 score header wi...
Render a recipe YAML as a compact visual overview with ASCII flow diagram and input table. Only invoke when the user exp...
Assemble a short-form 'vignette' ad from clean product cutouts composited over a kinetic background video — birefnet cut...
Generate images and videos via renderful.ai API (FLUX, Kling, Sora, WAN, etc.) with crypto payments. Use when the user w...
Diagnose and repair Operator memory load failures.
Surface past learnings relevant to the current task before starting work. Searches correction history, recalls past mist...
Runs open-source ML models via Replicate API for image generation, LLMs, and audio. Use when calling Stable Diffusion, L...
Run any Replicate model (image gen, audio, video) by version ID
Reference system prompt and tool schemas for Replit. Use when imitating, comparing, or analyzing Replit's prompting styl...
Indexes and documents a codebase for persistent Claude context with minimal token overhead. Use when asked to index a re...
Rigor Intake helper for README-first deep learning repo reproduction. Use when the task is specifically to scan a reposi...
Convert code repositories to formatted text for LLM analysis. Use when the user wants to send code to Gemini, Claude, or...
Guide for using Repomix - a powerful tool that packs entire repositories into single, AI-friendly files. Use when packag...
Package code repositories into AI-friendly files. Use this skill when packaging codebases for AI analysis, creating repo...
Repository packaging for AI/LLM analysis. Capabilities: pack repos into single files, generate AI-friendly context, code...
Package external repositories into AI-friendly files for deep analysis. ALWAYS use when the user shares a github.com URL...
Fix for `mcp__RepoPrompt__agent_run` subagents (Codex CLI) silently pausing with "Waiting For Input" when asked to write...
Submit feedback to the template repo as a GitHub issue (bug, feature request, or confusion)
尚书省 report 模式回奏格式规范，定义回奏必备字段与结构
高级报告生成专家，支持多格式输出、数据可视化和交互式报告生成。
將使用者的自然語言需求，轉換成「Codex 可直接執行」的工程指令 prompt。 輸出內容包含：背景、明確目標、可驗收條件、操作步驟、不可違反的限制、與交付格式。 適用於任何語言、任何 repo、任何專案階段。
Flag the open PR as ready for human review — request GitHub review from the two teammates who did not open it, then post...
Classify any user request into a structured format for prompt generation. Use this skill whenever you need to understand...
Patterns for classifying and routing game development requests to expert agents. Use when parsing vague requests like "i...
文档重排序服务（Reranker）。基于深度学习模型对检索候选结果进行细粒度相关性打分与重新排序，显著提升检索结果的精准度（Top-K 准确率）。
Reranking retrieved documents with cross-encoders and LLM rerankers. Cohere Rerank v3, Voyage rerank-2, BGE reranker, Co...
Reranking patterns for improving search precision. Use when implementing cross-encoder reranking, LLM-based relevance sc...
Autonomous goal-directed iteration loop. Define goal + metric + guard, iterate with specialist agents (perf-optimizer, s...
Comprehensive research grounded in web data with explicit citations. Use when you need multi-source synthesis—comparison...
Research agent for external documentation, best practices, and library APIs via CLI tools
Research agent for external documentation, best practices, and library APIs via MCP tools
Research Vercel AI SDK patterns, streaming, and multi-provider integration using Exa code search and Ref documentation
Use when researching a focused question on the web and returning a dated, source-linked evidence brief without editing f...
Use this skill whenever a researcher wants to test, validate, stress-test, or falsify a research idea or hypothesis — es...
Deep-research a topic end to end and produce a permanent, reusable knowledge asset: archive every primary source verbati...
Research Drizzle ORM patterns, queries, and PostgreSQL integration using Exa code search and Ref documentation
Retrieve existing results, ideas, examples, counterexamples, and corrections across a research campaign, sibling reposit...
Look up current research information using the Parallel Chat API (primary) or Perplexity sonar-pro-search (academic pape...
Generate the final research overview from the top-ranked hypotheses.
ML paper pipeline: experiment design to submission.
Build structured research prompts through interactive multi-step hearings using AskUserQuestion with selection-based opt...
Research React 19 patterns, hooks, and best practices using Exa code search and Ref documentation
Turn a vague research direction into a problem-anchored, elegant, frontier-aware, implementation-oriented method plan vi...
Synthesize a topic against Littlebird memory and current sources. Use for a concise what-changed research brief. Read RE...
Research similar solved problems and viable approaches on GitHub, Hugging Face, Zhihu, Xiaohongshu, official conference...
树状搜索科研迭代与三轮审查编排。多分支探索 idea/方法/模块/参数变体→执行实验→评分→剪枝→深化；编排三轮连带审查（拆子问题→分派→合并→收敛判定）。触发词：树状搜索、分支实验、迭代搜索、换模块跑迭代、三轮审查、tree search...
Proposal-first scientific writing pipeline, installed under the compatibility trigger researchwrite and the repository p...