AI & ML

oufinx Bundle

Розподіл дослідницької роботи між паралельними агентами - ексклюзивні зони джерел, друга вісь ракурсів, стелі викликів, і куди насправді згорає бюджет. Use when planning or running multi-agent research, splitting a research task across parallel subagents, deciding how many agents to launch and what each one gets, or when the user asks «розділи роботу між агентами», «запусти розвідку», "orchestrate parallel research agents", "how should I split this research".

reebeecka Skill

Prove a ticket is done (acceptance evidence, tests, dual-pass review, PR, ticket update). Use when the user says a ticket is done, asks to open a PR, or the agent is about to claim the work is complete.

temmax Bundle

Use when a feature or change needs a wave-ready implementation plan for parallel or multi-agent execution. Do not use to implement the plan.

temmax Bundle

Use when implementation work should be delegated, parallelized, or routed across Claude or Codex agents, especially when isolated worktrees and independent supervision are required. Do not use for single-agent work.

zephyr236 Bundle

openclaw-sdk 是 Python 包 (pip install openclaw-sdk)，用于连接 OpenClaw 自主 AI Agent 框架。 当用户提到以下任何一种情况时，必须激活此 skill： - 需要连接、管理或与 OpenClaw agent 交互 - 使用 OpenClaw SDK 进行 Python 开发 - 需要创建 agent、执行任务、管理渠道、定时任务、技能等 - 任何涉及 openclaw_sdk 的导入、安装、配置或使用 - openclaw, openclaw-sdk, OpenClaw SDK 相关的任何问题 即使用户没有明确说出 "skill"，只要涉及 SDK 使用就应该触发。 此 skill 不包含实现代码——所有详细 API 参考在 references/ 目录下。

zephyr236 Skill

Use when starting any task to check if available skills and MCPs are sufficient, and search for missing ones to recommend to the user

aerospike Bundle

Guides Aerospike application development with official clients for Java/Go/Python/C#/Node.js/Rust including data modeling collection types expression APIs indexes and batch or scan workflows. Use when users build or review client code or tune data access against a model that already exists. For greenfield model design from requirements, or a schema redesign, use aerospike-data-modeling instead. Redirect cluster deployment sizing XDR and backup questions to Aerospike Operations documentation. Core database only; not Aerospike Graph.

aerospike Bundle

Designs a new Aerospike data model from requirements, producing a schema guide and schema summary: record granularity, key design, bin structure, relationship and consolidation decisions, and index strategy. Use when no schema exists yet, when redesigning an existing model, or when reviewing a proposed schema for structural defects. This is design-time work that precedes code. For writing or reviewing client code against a model that already exists, or for client APIs, policies, CDT operations, and expression usage, use aerospike-development instead. Core database only; not Aerospike Graph.

anentrypoint Bundle

Applies WFGY (Wan Fa Gui Yi), a portable reasoning discipline for reducing drift and incoherence across multi-step agent work, adapted from onestardao/WFGY's core mechanism. Compares each step against the stated goal, considers more than one approach before committing to ambiguous or high-stakes decisions, and checkpoints before risky steps so it can revert and bounded-retry instead of compounding a mistake. Records durable lessons in this project so future sessions inherit them. Use when starting any non-trivial multi-step task, when a task risks losing track of its original goal over many steps, when facing a decision with real alternatives worth comparing, or when the agent notices it has contradicted itself or gone in circles and needs a disciplined way to recover.

asw101 Bundle

Delegate to GitHub Copilot CLI via its Agent Client Protocol server. Built on the official agent-client-protocol Python SDK (Pydantic-typed messages). Sibling of /copilot-cli (JSONL) and /copilot-sdk (Go).

asw101 Bundle

Delegate to GitHub Copilot CLI via its `--output-format json` stream. Python runner parses JSONL events into a structured transcript with full event replay. Default model gpt-5.6-sol + high reasoning + long_context (1M) window. Pass --model / --effort / --context to override; --multi to fan out across models.

ccancellieri Skill

Use when the user asks for a "comprehensive code review", "multi-model review", "ensemble review", "deep review", or "parallel model review". Dispatches three specialized Claude model tiers (Opus, Sonnet, Haiku) to review code in parallel, each focusing on different aspects. Aggregates findings into a unified severity-ranked report. Invoke when reviewing critical code paths, production deployments, architectural changes, or when the user wants thorough analysis beyond single-model review.

ccancellieri Skill

Use when the user asks to "compact memory", "create a master state snapshot", "summarize the project state", "restructure memory", or invokes /hub-compact-master-state, /role-senior-project-architect, or similar. Generates a 4-section snapshot (Architecture / Invariants / Active Working Set / Recent Pivots) for a multi-repo project from auto-memory deltas. Always uses the best configured model (tier_smart) and surfaces inferred assumptions for verification rather than presenting them as facts.

cnoles1980 Bundle

Pricing, packaging, and monetization strategy skill for product ideas, SaaS tools, AI products, information products, dashboards, APIs, newsletters, and emerging software businesses. Use when Codex should evaluate what to charge for, who should pay, how to package value, what pricing model to test, or whether a product should be monetized at all.

damien-xai Skill

Use the connected GitHub MCP for issues, pull requests, reviews, and repository data. Use when the user mentions GitHub, issues, PRs, reviews, branches, or runs /github.

epam Bundle

TODO: # Write a 2-3 sentence description of what this skill does and when the agent should use it. # Cover: what types of calculations it handles, what operations are supported, # and what user requests should trigger this skill (e.g. calculate, compute, evaluate, solve, etc.).

epam Bundle

TODO: Write when the agent should activate this skill. # Cover: what content types it handles and what user actions trigger it.

novelsavage Bundle

個人LLM Wiki（~/Projects/llm-wiki）をLLMの長期記憶として自動参照・維持する。ユーザー本人の経歴、経験、研究、制作物、人間関係、関心、文体、価値観、過去の作業や決定が現在の依頼に関係するとき、または会話・作業から今後も再利用できる恒久的な知見が確定したときに、明示的な指示がなくても使う。「以前の話」「自分について」「Wikiに反映」「覚えておいて」「Wikiを整理」や $wiki でも使う。

nvidia-bionemo Bundle

Agent runbook for Proteina-Complexa parameter sweeps through the separately installed first-party CLI and config generator. Use for cartesian-product hyperparameter scans, sweeper YAML authoring, configuration comparisons, ablations, tuning, success-rate ranking, or Pareto searches over quality and wall-clock cost. SKILL.md performs the orchestration; bundled executables only inspect local readiness and write a run manifest.

nvidia-bionemo Bundle

Agent runbook for orchestrating an end-to-end protein, ligand-binder, or AME design through the separately installed first-party Proteina-Complexa CLI. Use for requests such as "design a binder", "run complexa design", "de novo binder", ligand binding, motif scaffolding, beam search, FK steering, MCTS, refolding, or reporting design success and diversity. SKILL.md selects and invokes the project CLI; bundled executables only inspect local readiness and write a run manifest, and do not implement model inference themselves.

nvidia-bionemo Bundle

Agent runbook for evaluating an existing PDB directory through the separately installed first-party Proteina-Complexa CLI. Use to refold or score designs, compute interface pAE, i_pLDDT, scRMSD, motif RMSD, or designability, run complexa analysis, or assess third-party BindCraft, AlphaProteo, or RFdiffusion outputs. SKILL.md selects and invokes the evaluation configuration; bundled executables only inspect local readiness and write a run manifest.

oguzbilgic Skill

Add a Model Context Protocol (MCP) server to extend your toolbox with external tools

pnx2003 Bundle

Research, scope, design, execute, and report rigorous audits of AI agent systems. Use for Agent audit literature reviews, frontier-lab comparisons, alignment or hidden-objective audits, agent security red teams, AI-control and monitor evaluations, tool/MCP/permission audits, production observability and incident forensics, benchmark audits, governance assurance, audit plans, evidence registers, reproducibility reviews, or research-gap generation. Distinguish model evaluation from whole-system auditing and pre-deployment propensity testing from runtime control evidence.

pnx2003 Bundle

Research, classify, reproduce, and design recursive self-improvement (RSI) systems for LLM agents. Use for Agent RSI literature reviews, classic-paper reading lists, self-evolving/self-improving agent comparisons, repository audits, reproducibility plans, or experiments involving persistent memory, skills, prompts, tools, workflows, agent code, model weights, or the updater itself. Distinguishes durable recursive improvement from ordinary reflection, self-correction, and prompt iteration.

pnx2003 Bundle

Research, select, audit, reproduce, and design benchmark suites for LLM agents. Use for agent-benchmark surveys, related-work comparisons, evaluation plans, benchmark selection, leaderboard interpretation, reproducibility reviews, or experiments on web agents, computer-use agents, tool/API agents, software-engineering agents, deep-research/scientific agents, memory agents, multi-agent systems, safety/security agents, and embodied agents. Also use when comparing a base model, an agent scaffold/harness, or a complete agent system; when diagnosing saturation, contamination, reward hacking, environment drift, evaluator validity, cost, or reliability; and when studying model–harness co-evolution. Distinguish full interactive benchmarks from static diagnostics, environments, harnesses, and evaluator benchmarks.

pnx2003 Bundle

Design, audit, implement, and evaluate sampling strategies for LLM reasoning and interactive-agent rollouts. Use for rollout sampling, trajectory sampling, inference-time scaling, temperature/top-k/top-p/min-p, repeated sampling, pass@k, self-consistency, Best-of-N, verifier-guided search, beam/ToT/MCTS/DVTS, power sampling, MCMC/SMC, rejection sampling, importance sampling, off-policy correction, GRPO/PPO rollout design, rollout diversity, and train-time trajectory selection.

qqqqqf-q Skill

Drive real macOS applications through the CUA Driver MCP server when the user asks to inspect, operate, or automate visible desktop UI.

rodri-oliveira-dev Skill

Use esta skill para auditar qualidade de testes .NET neste repositorio, encontrando anti-padroes como asserts fracos, ausencia de asserts, flakiness, over-mocking, acoplamento a implementacao, dependencia de ordem, sleeps, dados magicos e cobertura artificial. Nao use para escrever testes novos do zero ou migrar framework.

rodri-oliveira-dev Skill

Use esta skill para revisar, refatorar ou melhorar codigo .NET/C# com foco em qualidade, legibilidade, manutencao, testabilidade, seguranca e performance. Nao use para reescrever codigo por preferencia estetica sem ganho tecnico claro.

solvapay Bundle

Use this skill when a user is choosing between SolvaPay surfaces rather than implementing one. Load for: any query that pits two options against each other ("checkout or SDK?", "MCP vs checkout?", "which one for my app?"), capability questions about SolvaPay, vague billing or monetization intent with no specific surface chosen, or assigned SolvaPay work without clear direction. Skip when the user has settled on exactly one surface and is asking to build it — single-surface intent plus implementation context (tech stack, command, or action verb) routes to that surface's dedicated skill instead.

solvapay Bundle

Use when the user is building a brand-new paid MCP server from zero — no existing server in play. The two entry points: (1) the user has an OpenAPI or Swagger spec they want converted into monetized MCP tools, or (2) they want to write the tools by hand on a fresh SolvaPay-wired scaffold. Also use for the `npm create solvapay` CLI command with `--type mcp`. This skill covers the full greenfield journey: spec parsing, tool authoring, Cloudflare Workers deployment, and SolvaPay paywall setup on a new codebase. Skip when the user already has a running MCP server and wants to add a paywall without rebuilding — use the existing-server skill. Skip for web checkout pages, Lovable flows, and SDK-only integrations.

viamagic333-lang Bundle

Check whether the user has done similar work before — across past agent chats, projects on disk, git history, and existing skills — and return a verdict: "делал" (done), "делал похоже" (similar), or "не делал" (never). Use automatically when a new task, request, or project is described, before starting work. Also use when the user invokes /FMSI explicitly, says "я делал это раньше?", "проверь делал ли я такое", "FMSI", or asks whether something has been done before.

viamagic333-lang Bundle

One-sentence summary of what this skill does and when to use it. Include trigger terms — words the user might say that should make the agent activate this skill. Use when the user says "X", "Y", "Z", or asks about <topic>.

vodoslopper Bundle

Build, rebuild, and troubleshoot ALT Linux RPM packages from Gear Git repositories using Gear and Hasher. Use when an agent needs to inspect a spec and `.gear/rules`, run the project's `gear --commit --hasher -- hsh-rebuild --no-sisyphus-check=packager,gpg` workflow, capture and analyze its build log, locate RPM artifacts, or fix Gear export, BuildRequires, RPM, compilation, file-list, or Sisyphus-check failures.

yaacovcorcos Bundle

Create or improve portable agent skills with clear trigger metadata, focused instructions, bundled resources, safety boundaries, and real validation. Use when writing, reviewing, refactoring, packaging, or publishing a SKILL.md skill for PapiSkill, Codex, Claude Code, Factory, GitHub Copilot, Cursor, VS Code, or another agent runtime.

yangaobo0235 Skill

Use when a user asks to install, start, or troubleshoot AI 应用经验上传助手; scan supported local Agent sessions; create an Assessment Scope; or submit assessment data.