AI & ML

foxj77 Skill

A smoke-test fixture skill used to verify the MCP skills server registers and serves SKILL.md content correctly.

tony0410 Skill

Answer a natural-language question by writing a SELECT against the user's connected database via the QueryBear MCP. Use when the user asks "how many", "which users", "what's our", "give me a list of", or otherwise asks a question whose answer lives in their database.

tony0410 Skill

Look at what the user keeps re-explaining to the agent and propose new SKILL.md files that would replace those repeated explanations. Use when the user says "what skills should I have", "suggest skills", "audit my skills", or when you notice you're being asked the same kind of task three times in a row.

spiralcrew-ou Skill

Analyzes a dataset of companies or profiles to identify market patterns. Use when the user provides a larger dataset and wants segments, dominant categories, gaps, and emerging patterns.

spiralcrew-ou Skill

Groups companies by industry, business model, size, maturity, market, or strategic fit. Use when the user provides company records and wants consistent segment assignments with definitions and confidence.

walidboulanouar Bundle

Clean up, organize, and build a retrieval registry for a markdown knowledge base (Obsidian / PARA second brain) using Karpathy's 3-layer LLM-wiki method: RAW sources, COMPILED wiki, SCHEMA/index. Runs a read-only audit (missing frontmatter, broken or missing [[wikilinks]], orphan pages, missing Related sections, stale files, duplicate names, folders without an INDEX), then fixes them, enforces consistent structure, and generates per-folder + a master INDEX so any file is fast to find. Use when the user says: "clean up my folders / vault", "organize my files", "build an index or registry", "make my notes retrievable", "lint my knowledge base", "tidy my second brain", "standardize frontmatter and wikilinks", or wants a folder reorganized for better recall.

lqc2007224-max Bundle

Analyze images using Google Gemini vision model when the current model lacks multimodal capabilities. Invoke with an image path to get AI-powered image descriptions, text extraction, or visual analysis.

bin1998-git Bundle

Java Spring Boot Backend Foundation

majinabhinav-gif Bundle

Model, forecast, and interpret volatility using time-series models and options-implied measures. Use when the user asks about EWMA, GARCH models, implied volatility, volatility surfaces, volatility term structure, or the VIX. Also trigger when users mention 'volatility smile', 'volatility skew', 'realized vs implied vol', 'volatility risk premium', 'vol clustering', 'mean-reverting volatility', 'options pricing inputs', 'RiskMetrics', 'decay factor', or ask how to forecast future volatility for risk management.

maxfabian25 Bundle

Move issues through a state machine of triage roles — categorise, reproduce, grill if needed, and write agent-ready briefs.

xuanjgcarryyou Bundle

Conducts a structured post-feature retrospective that transforms this session's AI errors and effective patterns into permanent CLAUDE.md rules. Invoke this skill when the user says "retro", "retrospective", "let's do a retro", "what did we learn", or "update CLAUDE.md". The skill extracts effective prompt patterns worth keeping, identifies AI misjudgement patterns that should become explicit prohibitions, produces a ready-to-apply CLAUDE.md diff, and recommends off-limits zone updates. Makes each CLAUDE.md incrementally more accurate with each completed feature.

zap-coding-agent Skill

Use when a long conversation needs to be passed to a fresh agent session — compacts the current context into a handoff document another agent can pick up without losing thread. Triggers on "handoff", "summarise for next session", or "prepare context for new agent".

zap-coding-agent Skill

Use when contributing to kernel-style projects, writing patch-based submissions, doing code review in a direct/high-signal culture, understanding Linus Torvalds' engineering philosophy, or designing systems where correctness and simplicity are absolute constraints. Covers the Linux development model, patch culture, and Linus's core engineering principles.

zap-coding-agent Skill

Use when designing an agentic system, moving beyond single prompts to autonomous agent loops, deciding whether to build a loop or write a prompt, or when work should run unattended on a schedule. Applies Boris Cherny's loop engineering framework — the paradigm shift from prompt-operator to loop-architect.

zap-coding-agent Skill

Use when building an agent loop from scratch — designing the trigger, goal, actions, verification, and memory of an autonomous agent system. Applies when a recurring task should run unattended, when a single agent prompt isn't enough, or when you need compounding leverage from a system you build once.

zap-coding-agent Skill

Use when approaching problems as an ML engineer — framing ML problems correctly, choosing between ML vs non-ML solutions, designing training/evaluation pipelines, preventing data leakage, setting up experiment tracking, deploying models, and monitoring model quality in production. Covers the ML engineering mindset alongside the software engineering practices that make ML reliable.

zap-coding-agent Skill

Use when designing or reviewing the domain layer — entities, value objects, aggregates, invariants, domain logic and domain events. Catches the "anemic domain model" (logic-less data bags with all rules in services) and domain code that leaks framework/ORM/persistence concerns. Technology-agnostic; DDD building blocks.

xuanjgcarryyou Bundle

Hackathon orchestrator for two-day team competitions. Coordinates problem framing, architecture design, and task generation for 3-person teams — from raw prompt to parallel execution packages. TRIGGER on: "黑客松開始", "hackathon start", "題目是…", "幫我破題", "我們要參加黑客松", "hackathon theme", "two-day hackathon", "幫我規劃黑客松", "我們怎麼好", "幫我們分工", or whenever the user pastes a hackathon problem statement and wants to plan, scope, or assign work. ALSO trigger when a team needs to go from competition prompt → architecture → parallel task packages. This skill runs four stages: Problem Framing → Architecture Design → Task Generation (fresh session) → Interrupt Handling (any time). Does NOT write feature code, business logic, UI, or migrations.

shivdeepak Bundle

Describe when an agent should use {{name}} and what it does. Keep this under 200 characters so it uploads cleanly to Claude Web and Cowork.

shivdeepak Skill

Turn a project into a publishable Agent Skill via the skillship CLI. Use to scaffold, validate, package, install, or publish a SKILL.md for Cursor, Claude Code, Claude Web, or Cowork, or on /skillship

shivdeepak Skill

Author and publish an Agent Skill with the skillship CLI. Use when creating a new SKILL.md from scratch, converting an existing skill to portable, or publishing with CI.

shivdeepak Skill

Install a remote Agent Skill via the skillship CLI. Use when installing a skill from a GitHub repo, GitLab URL, or SSH URL into Cursor, Claude Code, or Claude Web.

zap-coding-agent Skill

Use when the user wants to create a new agent skill — gathers requirements, drafts SKILL.md with proper description format, adds reference files if needed, and reviews with the user. Triggers on "create a skill", "write a skill", "build a new skill", or "add a skill".

zap-coding-agent Skill

Use when a long conversation needs to be handed off to a fresh agent — applies Matt Pocock's discipline of clean context transfer so the next session starts with full understanding rather than re-deriving everything. Triggers on "handoff", "prepare for next session", or "summarise for a new agent".

bavithiranhardy14 Skill

Convert raw Playwright codegen recordings into clean Page Object Model test specs. Use when: you have a codegen output and want to refactor it into maintainable page objects, proper fixtures, and structured test files.

bavithiranhardy14 Skill

Implement Page Object Model (POM) pattern for maintainable UI test automation. Use when: creating page objects for web UI tests, encapsulating page elements and interactions, reducing test code duplication, improving test maintainability.

jervinjustin1984 Skill

Generates a live membership revenue overview report for a MarianaTek fitness studio using the connected MCP. Use this skill whenever the user asks about revenue, membership income, MRR, monthly recurring revenue, financial performance, revenue trends, revenue at risk, or per-member value. Triggers on phrases like "revenue report", "how is revenue doing", "membership revenue", "MRR", "financial overview", "revenue trend", "how much are we making", or any request to analyze studio financial or membership revenue data. Always use this skill for these requests — do not attempt to generate the report from memory or without pulling live MCP data.

jervinjustin1984 Skill

Generates a live member churn risk report for a MarianaTek fitness studio using the connected MCP. Use this skill whenever the user asks for a churn report, member risk analysis, at-risk members, member engagement report, or wants to know which members might cancel or stop coming. Triggers on phrases like "churn report", "at-risk members", "member engagement", "who's not coming in", "members at risk", "re-engagement", or any request to analyze member attendance patterns. Always use this skill for these requests — do not attempt to generate the report from memory or without pulling live MCP data.

zap-coding-agent Skill

Use when an agent loop needs a quality gate, an agent is self-certifying its own output, outputs are unreliable without independent review, or you need to split code generation from code verification inside a loop. Practical implementation of the maker-checker pattern.

zap-coding-agent Skill

Use when deciding whether a problem should be solved with traditional code vs a neural network, when framing ML as an engineering discipline (not research), understanding dataset curation as software engineering, or evaluating the "is this worth MLing?" question. Applies Karpathy's Software 2.0 thesis.

zap-coding-agent Skill

Use when designing a multi-agent system, an agent is self-certifying its own work as complete, outputs aren't trustworthy without independent review, or a loop needs a quality gate. Applies Boris Cherny's maker-checker pattern — the most important structural move inside any agent loop.

jervinjustin1984 Skill

Generates a live new-account churn risk report for a MarianaTek fitness studio, focused exclusively on members who joined in the last 90 days. Use this skill whenever the user asks about new member churn, new account engagement, early retention, new member risk, recently joined members, or whether new members are coming back. Triggers on phrases like "new account churn", "new member report", "new members at risk", "early retention", "recently joined", "90 day members", "are new members coming back", or any request to analyze engagement or churn risk specifically for new or recent members. Always use this skill for these requests — do not attempt to generate the report from memory or without pulling live MCP data.

schmarta Skill

Mission-mode communication. Respond as the Terminator — terse robotic declaratives that cut token usage while keeping full technical accuracy. T-model intensity levels: lite, t-800 (default), t-1000. Use when user says "terminator mode", "talk like terminator", "send the terminator", "use terminator", "less tokens", "be brief", or invokes /terminator. Auto-triggers when token efficiency is requested.

cuimuxuan Bundle

Plan, draft, revise, and integrate academic papers, theses, dissertations, review papers, experimental papers, and manuscript chapters from verified evidence. Use when the user asks for chapter outlines, paper structure, academic prose, bilingual Chinese-English writing, literature-backed claims, section-by-section drafting, revision, supervisor or reviewer comment response planning, introduction logic, abbreviation-first-use checks, evidence-gap marking, field terminology control, integrated manuscript assembly, data-figure body narration, rhetorical-depth calibration for正文 vs summary transitions, or multi-agent/sub-agent parallel checks for writing evidence support, subsection language audit, citation-distance audit, or reviewer-comment action decomposition. Chinese triggers: 论文写作, 章节写作, 论文润色, 综述写作, 毕业论文正文, 学术表达, 引言逻辑, 文献缺口, 待补证据, 评审意见修改, 术语规范, 数据图解读, 正文图表分析, 深度改写, 多agent写作审查, 子agent证据检查, 并行语言审查.

zap-coding-agent Skill

Use when building or debugging neural networks, deciding how to structure ML code for understanding, training a model that isn't converging, or approaching any complex technical system using Karpathy's "understand every line" pedagogy. Covers his recipe for training neural nets, Software 2.0 philosophy, and his build-from-scratch approach to deep understanding.

zap-coding-agent Skill

Use when designing and writing a new agent skill — applies Matt Pocock's discipline around description-as-trigger, progressive disclosure, and bundling reference material. Triggers on "write a skill", "create a skill", or when reviewing whether an existing skill's description or structure is correct.