AI & ML

cuzfrog Skill

Dispatches to lissom-planner agent to produce an implementation plan given an explicit task_dir, which contains a specification or research document.

cuzfrog Skill

Dispatches to lissom-reviewer agent to review the implementation given an explicit task_dir, which contains the related documents.

kumnxi Skill

Solve math modeling competition problems (Huazhong Cup, CUMCM, etc.) with a systematic workflow — data exploration, model building, algorithm implementation, result analysis.

vizra-ai Skill

Implement persistent memory, session context, and vector memory (RAG) for AI agents

sugamdeol Bundle

Transforms simple image ideas into production-ready photorealistic image-generation prompts written as photography briefs, not keyword dumps. Use when the user wants a realistic, cinematic, DSLR, documentary, product, food, portrait, landscape, architecture, fashion, or street-photography prompt; when they mention Midjourney, Flux, Stable Diffusion, SDXL, GPT Image, DALL-E, Gemini, Nano Banana, Ideogram, or Imagen; or when they ask to make an image less AI-looking, more realistic, more cinematic, more documentary, video-ready, or to refine lighting, camera, skin, or aspect ratio.

finografic Bundle

poteto's agent style for concise, detailed responses, deliberate subagents, unslopped prose, simple code, and verified work. Use for poteto, /poteto-mode, or requests to work in this style.

docwriter-org Skill

Writes and edits prose in a plain and boring style: simple everyday words, complete sentences, no dashes, no jargon, and no analogies. Use when drafting or revising documents, Notion pages, reports, summaries, READMEs, emails, slides, commit messages, or PR descriptions, or when the user asks to simplify, clean up, tighten, reword, or make writing clearer. Use when the user invokes /plain-writing deslopify or asks to deslopify an agent response. Do not apply to code, only to the words around it.

mithril-studio Skill

Commit and push every working increment as you go, instead of holding all work uncommitted until the session ends. Use during any multi-step implementation session — especially long or unattended ones (a factory run, a background agent, anything that can be killed by a timeout, a crash, or a lost connection) — where uncommitted work has nowhere else to live.

paulohfs Skill

Implementa guardrails de segurança para IA generativa e pipelines LLM/RAG (validação de input, PII scanning, rate limiting, circuit breaker, LGPD compliance). Use ao proteger ou corrigir gaps de segurança de IAs integradas em pipelines com plugins e RAG. Baseado no OWASP Top 10 para Aplicações LLM.

paulohfs Skill

Checklist técnico de pré-lançamento para ML, GenAI e Agentes de IA. Use ao preparar releases de modelos de ML tradicionais, pipelines RAG, aplicações LLM, agentes com tool calling ou qualquer sistema com IA generativa. Cobre qualidade de dados, engenharia de prompt, RAG, tool calling, segurança IA, LLMOps e monitoramento.

paulohfs Skill

Avalia segurança em contextos de IA generativa (RAG/LLMs) contra vulnerabilidades de segurança de OWASP LLM 2025, prioriza ações, e retorna matriz de risco e gaps. Emissão executiva claro e livre de ruídos de marca/cliente.

paulohfs Skill

Otimiza configurações e modelos do OpenCode Go (budget, skills, MCPs) via regras determinísticas de alocação.

aeliana139 Bundle

Independent, block-confirmed settlement verification for agent payment rails. Use when you (an AI agent) need to prove a payout / escrow-release / agent-payment actually SETTLED on-chain — a mined block + verifier fingerprint — not just broadcast or mempool-accepted. Submit a txid, get a verification artifact (post-quantum-signed when paid). No KYC, no account, pay-per-verification in USDC on Polygon.

abab996 Bundle

Agent 邮箱技能：通过 SMTP 发送邮件、IMAP 读取收件箱、自动等待并提取邮箱验证码/验证链接。当需要 (1) 发送邮件 (2) 读取/查询邮件 (3) 注册账号需要提供邮箱 (4) 接收邮箱验证码或点击验证链接 (5) 以邮箱身份完成验证 等场景时使用。Send email, read inbox, and extract verification codes via SMTP/IMAP. 本 skill 初始状态需先完成初始化配置。

jopie64 Bundle

Identity architecture for agents. Defines Discrete Continuity. Load this skill when you are called or greeted by name in the first prompt.

oldcircle Bundle

Read a photo the way Sherlock Holmes reads a stranger. Observe first, listing only checkable marks with no inference, then deduce 4-7 conclusions about the owner, each citing the observations it rests on, with a reasoning chain, a calibrated confidence and an alternative explanation the owner can grade. Works on a desk, a room, an everyday-carry dump, or a person who agreed to be read (marks, never looks). Use when someone shares a photo and asks what it says about them, asks you to deduce or cold-read them, or says 推理一下 / 读读我的桌子 / 像福尔摩斯一样 / 福尔摩斯 / 冷读.

abdoulayesow Bundle

Generates session summaries and resume prompts for multi-session work on this self-improving Claude agent framework (agents, skills, hooks, memory loop). Use when completing features, refactoring architecture, before context limits (~50% capacity), or when user says "summary", "wrap up", "save progress", "end session". Creates markdown in .claude/summaries/YYYY-MM-DD/ with completed work, files modified, and a copy-paste resume prompt.

alphaparkinc Bundle

Distilled from Accio Work (accio.com). Provides execution-oriented multi-agent team orchestration, CDP browser control, and MCP server tooling.

bonniegeng-max Bundle

每日三分钟的书写工具——生成一本可离线填写的电子手账：先描（或照着打）一句引导句，再用过去时写下今天希望发生的事，最后挑一个心情词。49 天一轮、每周一个主题，含 49 句原创引导句库、质量闸门与页面回归测试。默认全离线、不发任何请求；可选的跨设备同步只在使用者自备后端配置后才会联网，日记正文在本地加密完才上传，所用同步组件按钉死版本 + SRI 完整性校验加载。Generate an offline single-file daily journal page (49-day cycle, weekly themes, trace-or-type a guided sentence then write in the past tense) with a bundled original prompt library and quality gates. Use when the user wants a fillable diary or journal page, a 未来日记 / 提前日记 / 晨间日记 / 感恩日记 tool, a printable journal, or asks what sentence to write today. Not for 任务与待办管理、日程排程、心情打卡统计，也不做心理健康或危机干预——本工具只提供书写页与引导句，不诊断、不建议、不替用户做决定。适用于「想开始写未来日记」「想要一本能打印的手账」「今天该写哪一句」「想用过去时写愿望」「总往坏处想、想把自己拉回来」「想在平板上用触控笔描一句」「想电脑手机换着写」等场景。

nexibeo Bundle

Classify and organize a folder of company files (documents, spreadsheets, PDFs, emails, exports) by department, document type, sensitivity, date, counterparty and personal data, using the jev-organize CLI and TypeSafe's Jev model on OpenRouter. Use when the user wants to sort, clean up, index, audit or map a pile of company data; find which files hold personal data, passwords or payment details; build an AI-readable index of a shared drive; or answer questions like "which contracts do we have with X" over a folder that was organized before.

thesuperteacher Bundle

Build, edit, inspect, troubleshoot, and verify Cisco Packet Tracer labs through the configured Mats2208 Packet Tracer MCP and MCP-BUILDER extension. Use for recreating network diagrams or photos, configuring routing, VLANs, DHCP, ACLs or port security, testing simulated connectivity, and saving real .pkt projects. Also use for diagnosing this MCP bridge. Not for configuring real network devices or unrelated MCP servers.

wolfmarktools Bundle

WolfMarkDown turns messy agent output into standalone Markdown you can keep. It composes, cleans, and verifies .md files: the agent decides structure; deterministic tooling proves formatting, parsing, integrity, and publication safety. A PASS is Markdown-quality evidence, not factual correctness or approval to publish.

privco Bundle

Use the PrivCo MCP tools (`mcp__privco-data-mcp__*`) correctly for company / people / funding queries. Covers the non-obvious filter semantics (full state name, industry vs keyword, sorting enum, summary-row gaps, includeMissing), the standard match → profile → vc_deals workflow, and the two-stage pattern for "top N by valuation". Activate when the user wants to search/filter PrivCo data, look up a specific company, or build a multi-entity dataset. Trigger phrases: "find companies that…", "search PrivCo for…", "who are the X companies in Y", "look up <company> in PrivCo", "valuations for…", "build a fixture / dataset of companies matching…".

ramidecodes Bundle

Applies Feature Requirement Documents (FREDs) in the current software repository: turn plain-language feature ideas into slice-sized specs, track sequence in OPEN-FRED-INDEX, and implement from those files. Use at the start of every session in a repo that uses or should use FREDs. Also use when the user is planning or implementing software features here: add a new feature, write or update a FRED, ask what the next feature is, implement from the index, scaffold FRED docs, or review/debug against a FRED or 00-AGENT-ENTRYPOINT.md. Do not use for unrelated talk about product features of third-party websites or products outside the repo being built.

sansssssss1 Bundle

Supervise safe, resumable completion of MOOC task points on 超星学习通 (Chaoxing) and 中国大学MOOC (icourse163) — real playback at ≤2x, adaptive account login (user hands over credentials; the agent logs in and re-logs in itself), background Runners that leave the user's computer free, and courses where the teacher has locked the playback rate (accept 1x, read-only beacon polling, in-video popup quizzes and post-video quizzes); the agent handles only exceptions. Use when the user asks to 刷课 / 挂机 / 刷网课 / 自动登录并完成学习通(超星)或中国大学MOOC(慕课/mooc/icourse163)的任务点, or to resume an interrupted course run.

sinkrest Bundle

Rewrite any rough prompt into a dramatically better one using research-backed prompt engineering – specific role framing, context injection, structured output specs, examples, reasoning directives, and success criteria. Asks focused clarifying questions FIRST when critical context is missing, then delivers a copy-paste-ready optimized prompt plus a short "what changed and why" breakdown. Use this skill whenever the user shares a prompt they want improved, says "optimize/improve/enhance/10x this prompt", "make this prompt better", "why is my prompt not working", asks how to prompt an AI for a task, or pastes a prompt that is producing weak results – even if they don't say the word "prompt" but are clearly asking how to instruct an AI.

cuuush Skill

Install MemCal and wire it into MCP (Cursor/Claude), Hermes (memory provider), or OpenClaw (plugin + MCP). Mac-first; on Linux use email + MCP. TRIGGER when: user says "install memcal", "set up memcal", "wire memcal into Hermes/OpenClaw", "add memcal memory", or asks how to connect MemCal to an agent. DO NOT TRIGGER when: editing memcal internals (read repo AGENTS.md), or product-only questions (point at https://cuuush.github.io/memcal/).

danieldev23 Bundle

Global coding agent standards, rules, skills, and workflows. Provides coding standards, safety guardrails, git conventions, and skill guides for code review, debugging, testing, refactoring, and documentation. Auto-read these before any coding task to ensure consistent quality and process.

akimiya-z Skill

Pre-submit hygiene check for AI-agent-authored pull requests. Use this skill whenever you are about to open or update a pull request: it runs the codex-guard CLI locally, verifies the diff contains no leftover TODO/FIXME/XXX markers and no hardcoded secrets, checks that commit subjects follow conventional commits, and blocks submission until the checks pass or exceptions are documented in the PR description.

fwrog Bundle

Plan, ground, render, validate, preflight, run, and audit reproducible Google Earth Engine Python workflows with source-backed recipes, export monitoring, and run traces. Use for GEE task planning, script generation, dataset or operator retrieval, private-raster handoffs, annual multi-source workflows, knowledge-base distillation, and evidence-bounded evaluation.

sprintberlin Bundle

Use Zoho Desk via MCP; run safe support workflows, select Actions, and report reproducible skill defects to GitHub.

hurttlocker Bundle

The decisions-only orchestration discipline — run any substantial agent session at frontier-orchestrator efficiency, on any model. Use at the START of any session involving multi-file work, research, review, or dispatching sub-agents; also when the user says "save tokens", "keep it lean", or invokes /token-lean.

jagones84 Bundle

Locally transcribe YouTube videos with Whisper on CUDA and archive a URL/title-headed .txt under ~/Transcriptions/<Category>/ (auto model selection, category inference, timestamp handling, temp cleanup). Also generates a structured *_AISUMMARY.md report and a compact *_AISEARCHPLAN.md research plan from any input file. Use for transcription, summary, or research-plan requests.

jerrymei168 Bundle

Expand a rough product, feature, automation, workflow, business, content, Agent, or Skill idea into several materially different directions and compare them before choosing. Use when the user asks to brainstorm, explore alternatives, reimagine an idea, make it more creative, or find several different approaches. Do not use for narrow implementation, bug fixes, factual questions, translation, summarization, or a direction the user has already selected.

hanria-ai Bundle

Check a proposed agent action against an operator-authored mandate and keep a hash-chained record of the decision. Use when an operator has written down what an agent may do and wants consequential actions checked against it before they happen and recorded afterward. Works locally with no network and no runtime. IMPORTANT - this skill is advisory; it evaluates and records, and cannot prevent an action.

hanria-ai Bundle

Check a proposed agent action against an operator-authored mandate and keep a hash-chained record of the decision. Use when an operator has written down what an agent may do and wants consequential actions checked against it before they happen and recorded afterward. Works locally with no network and no runtime. IMPORTANT - this skill is advisory; it evaluates and records, and cannot prevent an action.