MLOps
模型部署、评估与运维
Knowledge Pipeline Skill
drshailesh88
```
compound-catalog
James-Traina
Capture solved research problems as categorized documentation with YAML frontmatter for fast lookup. Use when a methodological, estimation, or data problem has been solved and should be documented for future reference. Triggers on "that worked", "it's fixed", "problem solved", "estimation converges now", "proof is complete", "pipeline runs", or when /workflows:compound invokes solution documentation. Also triggered by "document this solution", "save this fix", "log this resolution".
multi-model-writer
drshailesh88
"Unified writing system with intelligent model routing. Default: Claude. Options: GLM-4.7 (cheapest), GPT-4o/mini, Gemini, Grok. Includes browser automation for web interfaces. Cost-aware routing based on task complexity."
cheap-model-testing
almeidamarcell
When working on any application that integrates with LLMs or pay-per-usage APIs, always use the cheapest available model during development and testing. Remind to upgrade to a production model before deployment.
audio-to-text
lucas-acc
Transcribe audio files to text with automatic language detection (supports Chinese and English)
MCP 工具集成专家
valkryhx
"集成 Model Context Protocol (MCP) 工具，通过 HTTP 连接与外部 MCP 服务交互，提供参数管理、工单生成等功能。"
参数 MCP 工具集成专家
valkryhx
"集成参数 Model Context Protocol (MCP) 工具，通过 HTTP 连接与外部参数 MCP 服务交互，提供参数管理、工单生成等功能。"
x-algo-pipeline
CloudAI-X
Explain the complete X recommendation algorithm pipeline. Use when users ask how posts are ranked, how the algorithm works, or want an overview of the recommendation system.
view-refactor
patrickserrano
Refactor SwiftUI view files for consistent structure, dependency injection, and Observation usage. Use when asked to clean up a SwiftUI view's layout, handle view models safely, or standardize how dependencies are initialized and passed.
research
sourabhJainR
Conduct bounded engineering research using repository evidence, external sources when available, explicit unknowns and a reusable evidence package.
modelslab-model-discovery
ModelsLab
Search and discover 50,000+ AI models on ModelsLab, check usage analytics, and monitor generation history via the Agent Control Plane API.
model-thinking
kcchien
Mental models toolkit for clearer thinking, better decisions, and problem-solving. Use when users face complex problems, need decision support, want to analyze situations from multiple angles, organize information, understand systems, predict outcomes, or learn about specific mental models. Triggers include phrases like "help me think through", "analyze this problem", "what models apply here", "how should I decide", "evaluate options", or direct model references (e.g., "use second-order thinking", "apply inversion"). 中文觸發：「思維模型」、「幫我分析」、「決策分析」、「多角度思考」、「怎麼判斷」、「幫我想清楚」、「系統思考」、「風險評估」。
thermal-assessment
LunCoSim
Perform detailed thermal analysis for spacecraft and rovers across various celestial environments. Use this skill to calculate heat balance, size radiators, determine MLI requirements, predict nodal temperatures, and plan thermal hardware. Trigger this for "thermal modeling," "radiator sizing," "temperature prediction," "heater sizing," or "thermal control."
optimize-prompt
jbrukh
Optimize a prompt through a critique-compress pipeline with semantic equivalence verification at each stage. Applies think-critically to improve the prompt, then compress-prompt to reduce it, validating that behavior is preserved after each transformation.
cost-modeling
LunCoSim
Perform parametric cost estimation and ROM (Rough Order of Magnitude) costing for space missions. Use this skill to generate lifecycle cost estimates using CERs, compare launch costs, estimate operations budgets, and support cost-driven trade studies. Trigger this for "cost estimate," "ROM cost," "mission cost," "launch cost," "cost model," "lifecycle cost," "operations cost," or whenever a trade study needs cost as a Figure of Merit.
write-article
jbrukh
Write accessible, engaging op-ed-style articles in the Brukhman voice, authoritative but conversational, technically grounded, with progressive narrative structure.
cost-estimation
LunCoSim
Perform parametric cost estimation and cost-risk analysis for space missions. Use this skill to generate ROM (Rough Order of Magnitude) costs, build Cost Estimating Relationships (CERs), produce cost breakdowns by subsystem and WBS element, and generate cost S-curves for risk assessment. Trigger this for "mission cost," "cost estimate," "ROM cost," "cost model," "cost-to-complete," "budget estimate," "proposal cost," "cost breakdown," or "cost risk."
apple-foundation-models
ios-agent
Use this skill when working with Apple's Foundation Models framework for on-device AI and LLM capabilities in iOS/macOS apps
prompt-engineering
maragudk
"Use this skill when crafting, reviewing, or improving prompts for LLM pipelines — including task prompts, system prompts, and LLM-as-Judge prompts. Triggers include: requests to write or refine a prompt, diagnose why an LLM produces inconsistent or incorrect outputs, bridge the gap between intent and model behavior, reduce ambiguity in instructions, add few-shot examples, structure complex prompts, or improve output formatting. Also use when the user needs help distinguishing specification failures (unclear instructions) from generalization failures (model limitations), or when iterating on prompts based on observed failure modes. Do NOT use for general coding tasks, document creation, or non-LLM writing."
swiftui
ios-agent
SwiftUI framework for building user interfaces
movement-science-stats
jjodx
Statistical quality checklist for movement science and neuroscience research. Auto-triggers when analyzing data, interpreting results, running statistics, writing results sections, or reviewing analysis code. Based on Makin & Orban de Xivry (2019, eLife).
llm-as-a-judge
maragudk
Build, validate, and deploy LLM-as-Judge evaluators for automated quality assessment of LLM pipeline outputs. Use this skill whenever the user wants to: create an automated evaluator for subjective or nuanced failure modes, write a judge prompt for Pass/Fail assessment, split labeled data for judge development, measure judge alignment (TPR/TNR), estimate true success rates with bias correction, or set up CI evaluation pipelines. Also trigger when the user mentions "judge prompt", "automated eval", "LLM evaluator", "grading prompt", "alignment metrics", "true positive rate", or wants to move from manual trace review to automated evaluation. This skill covers the full lifecycle: prompt design → data splitting → iterative refinement → success rate estimation.
rails-model-patterns
ag0os
ActiveRecord model patterns and conventions for Rails. Automatically invoked when working with models, associations, validations, scopes, callbacks, or database schema design. Triggers on "model", "ActiveRecord", "association", "has_many", "belongs_to", "validation", "validates", "scope", "callback", "migration", "schema", "index", "foreign key".
event-modeling
jwilger
Event modeling facilitation for discovering and designing event-sourced systems. Four phases: domain discovery, workflow design (9-step process), GWT scenario generation, and model validation. Activate when starting a new project, designing features, modeling domains, writing Given/When/Then scenarios, or discussing event sourcing and domain-driven design.