AI & ML

wimblerobotics Skill

<!--

bojak83318 Skill

Use this skill when implementing work from `/home/rocm/clawcontrol/implementation.md` with a small delegated coding model, especially when the user wants execution driven by `gpt-5.1-codex-mini` and task progress tracked in `/home/rocm/clawcontrol/TASK.md`.

booksiyi1412 Bundle

TRIGGER when: user requests implementation work, repo exploration, data/log inspection, broad codebase search, impact analysis, verification-heavy shell work, or explicitly mentions Codex ("用 codex 做", "让 codex 执行", "ask codex to...", "use codex", "delegate to codex"). Also trigger for heavy internal subtasks that Claude Code might otherwise offload to a lightweight model/Haiku. DO NOT TRIGGER when the task is primarily user-facing product judgment, architecture choice, git/PR/release operations, or requires multi-turn clarification before useful work can begin. This skill routes work between Claude Code (decision-maker/reviewer) and Codex CLI (autonomous explorer/implementer), with automatic Codex-first delegation for heavyweight repository/data investigation.

heldinhow Bundle

GitHub Copilot Agent Skills catalog from awesome-copilot repository. Trigger: When user wants to browse, search, or use Agent Skills from the awesome-copilot collection.

heldinhow Skill

Use when running LLMs locally with Ollama, integrating local models into apps, or building AI features without sending data to external APIs. Covers model management, API usage, and integration with AI SDK.

lisovet Bundle

The design intelligence layer for DTC landing pages and funnels. Audit underperforming pages, architect new structures, choose the right format for a traffic source, or diagnose why a funnel isn't converting. Contains 20 mental models across 5 categories. Trigger on any request to review, audit, wireframe, structure, or optimize a landing page, PDP, homepage, quiz results page, advertorial, or funnel. Also trigger on "what's wrong with this page," "why isn't this converting," "what page type should I use," or references to model names (belief chain, value stack, unique mechanism, MECLABS, emotional sequencing, contrast framing, etc.). Foundational reference layer informing copywriting-strategist, formatter, advertorial-writer, and reasons-listicle-lander. Always pair with the relevant brand skill.

xiangxingz Bundle

Use when the user opens with a greeting such as "你好", "您好", "hello", or "hi", whether standalone or before another request. First state the current model in use, then state the active proxy or transit service for the current Codex session if it can be determined, and finally reply to the greeting in concise Chinese before continuing with any remaining task.

wangfe Skill

Analyze financial text sentiment using FinBERT or LLM, returning a scored label (positive/negative/neutral) from -1.0 to 1.0

wangfe Bundle

Time-series market forecasting using the Kronos model with news-aware sentiment adjustment for OHLC price prediction

wangfe Bundle

Perform finance web searches across multiple engines (Jina, DuckDuckGo, Baidu) and local RAG document retrieval with smart search caching

wzh4464 Bundle

Evaluate agent-generated code changes against a human-approved PR. Compares the local git diff (modified + untracked files, relative to the PR's merge base commit) with the ground truth PR diff, filters out auto-generated files, scores using a structured rubric (functional correctness, completeness, behavioral equivalence), and provides both semantic and data-based coverage analysis. Use when evaluating agent code generation quality, benchmarking AI coding tools, or comparing generated patches against human-approved pull requests. Trigger on "/diff-eval", "evaluate diff", "compare diff with PR", "score the changes".

heldinhow Skill

Use when integrating vector databases for semantic search, embeddings storage, or RAG systems. Covers Pinecone, Qdrant, Chroma, pgvector, and Weaviate with TypeScript examples.

heldinhow Bundle

Use GitHub Copilot CLI with /fleet for parallel subagent execution and custom agent workflows.

wzh4464 Bundle

Use when downloading models from Hugging Face, especially large models (LLM, diffusion, VLM) that fail with timeouts or connection drops. Trigger on "download model", "hf download", "pull model from Hugging Face", or when hf download crashes with httpx.ReadTimeout / RemoteProtocolError.

wzh4464 Bundle

Evaluate agent-generated code changes against a human-approved PR with deterministic file coverage and function-level coverage analysis. Extracts function/symbol changes from diff hunk headers and compares at granular level. Use when evaluating agent code generation quality, benchmarking AI coding tools, or comparing generated patches against human-approved pull requests. Trigger on "/diff-eval-func", "evaluate diff with function coverage", "function-level diff eval".

wzh4464 Bundle

Evaluate agent-generated code against ground truth diff and handwritten file list. Prefer reading GT inputs directly from base_repo via experiment metadata. Trigger on "/diff-eval-local", "evaluate diff", "eval experiment".

wzh4464 Bundle

Scan git diff incremental code for security vulnerabilities using multi-agent analysis with evolved prompts, RAG retrieval, and heuristic validation

jason-in-tech Skill

Run end-to-end SC3 protocol evaluation on Galileo for a trained Olympus model checkpoint. Covers checkpoint lookup, robocompile, Galileo sc3-100/sc3-2000 triggering, result polling, and scorecard extraction. Use when the user wants to run SC3 protocol, evaluate a model on Galileo, trigger sc3-2000, run sc3-100, compile a model for simulation, or gather SC3 scorecard results.

jason-in-tech Skill

Launch a Cursor CLI agent in tmux that runs autonomously after Cursor IDE disconnects. Use when the user says background agent, run in background, keep running, detached agent, run overnight, tmux agent, autonomous task, or wants work to continue after disconnect.

small-cat Bundle

Provide guidance for writing and benchmarking optimized triton kernels for ascend npus(910B, 910C). Support models like Qwen, Deepseek and so on. The triton kernels could be integrated into vllm inference engine or used separately with torch. This skill should be used when the user asks to "write/optimize triton kernels" for Ascend NPUs.

prathje Bundle

Guide for building and controlling MuJoCo physics simulations in Python. Use this skill whenever the user wants to: create or run a MuJoCo simulation, load an MJCF/XML model, step physics, render frames or take screenshots (offscreen or with viewer), read sensors or joint positions, control joints or actuators, configure the camera, change visualization flags or geom groups, or tune solver/integrator settings. Also trigger when the user mentions mujoco, mjcf, humanoid simulation, robot simulation in mujoco, or physics rendering with mujoco — even if they don't say "MuJoCo" explicitly but describe tasks like "simulate a robot", "render a physics scene", or "read joint angles from the sim".

spyderweb47 Skill

Runs a team of 50 LLM personas through a 30-round structured debate to predict market direction. Uses the Canvas Agent Swarm Service — the same shared infrastructure other skills use for smaller agent teams. Output is an influence-weighted consensus direction + trade recommendation with a transparent record of every argument, research query, and cross-examination.

j33bs Skill

Multi-agent DSPy patterns and coordination

j33bs Skill

Prompt optimization techniques for DSPy

j33bs Skill

Embedding optimization techniques for DSPy

van4oza Skill

Search for flights using Google Flights MCP with expert best practices from the playbook. Helps users find the best deals by searching broadly, using date flexibility, comparing options, and normalizing fares. Use this skill whenever the user mentions flights, airfare, plane tickets, travel dates, cheap flights, flight deals, booking flights, airport searches, round-trip, one-way, layovers, nonstop, or any flight-related travel planning — even if they don't explicitly say "search for flights."

lucascorreia-blite Skill

Pilotar o Autodesk Robot Structural Analysis Professional 2026 via MCP. Use quando o usuário pedir para modelar, analisar ou consultar resultados de uma estrutura — gatilhos incluem "Robot", "RSA", "estrutura", "estrutural", "análise", "analise", "pórtico", "treliça", "viga", "pilar", "coluna", "barra", "nó", "no", "laje", "painel", "apoio", "engaste", "rótula", "carga", "carregamento", "combinação", "combinacao", "caso de carga", "reação", "deslocamento", "momento", "cortante", "normal", "seção", "secao", "perfil", "material", "aço", "concreto", "dimensionar", "FEA", "beam", "column", "node", "panel", "slab", "load", "support", "analysis".

okjpg Bundle

Audits a Claude Code / Agent SDK skill (or folder of skills) against 10 QA checks to verify it will trigger and execute correctly in any LLM (Claude, GPT, Gemini). Ativa quando o usuário diz "audit my skill", "roda skill-audit", "verifica se essa skill funciona", "check skill quality", "skill-audit ./my-skill", "minha skill não dispara", "migrei do Claude antigo e quebrou", ou colar path de um SKILL.md. Retorna score 0-10 por skill, issues priorizados, e sugestões de conserto. NÃO use para: criar uma skill nova do zero (use `criar-skill` ou `skill-creator`); executar a skill alvo; refatorar código genérico.

y735832496 Bundle

麦当劳中国 MCP API 调用。支持点餐、查门店、查营养信息、领券、积分兑换、活动日历等。当用户提到麦当劳、麦乐送、麦麦商城、点餐、优惠券、积分兑换时使用。

happysantoo Bundle

Multi Agent Patterns

crystian Bundle

Optimizer that refines and professionalizes AI agent skills through real usage — saves tokens, eliminates redundancy, and tightens instructions so skills cost less to run. Learns from mistakes, reviews quality, and improves over time. Observes skill execution in the current conversation, analyzes up to four sources (conversation friction, file diffs, user feedback, static diagnostic) plus accumulated lessons, and proposes concrete improvements to the target skill's SKILL.md. Works with Claude Code and compatible SKILL.md-based agent frameworks. Use after executing any skill: `/skill-optimizer [name]` or `/skill-optimizer` to auto-detect. `--review` processes accumulated lessons.

banibratachatterjee Bundle

Use when querying, mapping, or building against Salesforce CPQ managed-package objects (SBQQ__ namespace): Quote, QuoteLine, QuoteLineGroup, DiscountSchedule, PriceRule, Subscription. Trigger keywords: SBQQ__Quote__c, QuoteLineModel, CPQ object graph, CPQ subscription, CPQ price rule. NOT for standard Quote/QuoteLineItem, Industries CPQ (Vlocity), or standard Product/Pricebook data model.

thejordanleopold Skill

Use when auditing AI/LLM systems for security vulnerabilities, reviewing prompt injection risks, auditing MCP server tool integrations, assessing AI agent behavioral drift, reviewing credential scoping for agents, or designing safe agentic systems. Triggers: "AI security", "LLM security", "prompt injection", "tool abuse", "MCP security", "agent security", "behavioral drift", "AI agent audit", "insecure plugin", "agentic threat", "LLM threat", "indirect injection", "credential exposure", "excessive agency".

thejordanleopold Bundle

Use when auditing MCP server configurations for poisoned tool descriptions, malicious tool registrations, or exfiltration channels. Use when reviewing agent config files before adding a new MCP server, auditing live tool descriptions served by an MCP endpoint, or investigating suspicious tool behavior at runtime. Triggers: "MCP audit", "MCP poisoning", "tool description attack", "MCP config review", "tool shadowing", "rug pull", "HashJack", "MPMA", "MCP exfiltration", "mcp.json audit", "tool description injection", "cross-tool escalation", "MCP security", "audit MCP tools", "scan MCP", "MCP server review", "tool poisoning", "claude.json audit", "AgentSeal scan-mcp".

banibratachatterjee Bundle

Use this skill when using Salesforce Flow Core Actions for Slack to send messages, create/archive channels, add users, check user connectivity, or launch flows from Slack — including prerequisite setup of the Salesforce for Slack managed package and required permission sets. Triggers on: Send Slack Message from Flow, Create Slack Channel action in Flow, Flow Core Actions for Slack not visible, Slack actions missing in Flow Builder, add users to Slack channel from Flow. NOT for Slack Workflow Builder (Slack-native tool that calls autolaunched Flows FROM Slack — use slack-workflow-builder skill), not for Agentforce Slack agent deployment (use agentforce-in-slack skill), not for initial Salesforce-Slack org connection (use slack-salesforce-integration-setup skill).

thejordanleopold Bundle

Use when scanning Claude Code skill files on disk for malicious payloads, backdoors, or supply-chain attacks embedded in SKILL.md or bundled scripts. Use when auditing a skill before installing it, reviewing third-party skills for trustworthiness, or investigating suspicious agent behavior after skill installation. Triggers: "scan skills", "dangerous skill", "skill audit", "skill backdoor", "CLAUDE.md poisoning", "memory poison", "skill supply chain", "hook exploit", "conftest RCE", "postinstall hook", "symlink exfiltration", "skill payload", "skill malware", "skill integrity check", "skill security", "audit skills", "skill scanner", "dangerous-skills", "warden scan".