Machine Learning Skills

| Apify integration. Manage Actors, Datasets, KeyValueStores, RequestQueues, Tasks. Use when the user wants to interact with Apify data.

| Benchmark Email integration. Manage data, records, and automate workflows. Use when the user wants to interact with Benchmark Email data.

| Cohere integration. Manage Documents, Models, Datasets, Jobs. Use when the user wants to interact with Cohere data.

| Geckoboard integration. Manage Dashboards, Datasets. Use when the user wants to interact with Geckoboard data.

| Gpt-trainer integration. Manage Users, Roles, Goals, Pipelines, Filters, Organizations. Use when the user wants to interact with Gpt-trainer data.

| Hugging Face integration. Manage Models, Datasets, Spaces. Use when the user wants to interact with Hugging Face data.

| Metatext.AI Inference API integration. Manage data, records, and automate workflows. Use when the user wants to interact with Metatext.AI Inference API data.

| Nasdaq Data Link integration. Manage Datasets. Use when the user wants to interact with Nasdaq Data Link data.

Build and govern a 50-200 example domain-specific held-out benchmark sampled from real traffic. Distinct from public benchmarks (MMLU/HumanEval/GSM8K via lm-evaluation-harness) which measure GENERAL capability. Only a held-out domain set predicts whether THIS system works on YOUR data. Collect real examples, label, hold out (never train/prompt on it), size 50-200, version it, refresh on drift.

>- into writable context) under ~25k tokens, separate "read" from "edit", delegate breadth to a read-only repo-map, and drop files once edited. Use when an LLM coder-agent edits multiple files, when the working set must stay focused, or when the model starts editing the wrong file / missing targets because too much context dilutes attention. Search working file budget, context dilution, lost in the middle.

SOP for building LLM applications on Dify — visual workflow + chatflow + agent + RAG knowledge base + plugin marketplace + observability, self-hostable. Use when shipping LLM apps fast with a "no-code to pro-code" gradient, especially when non-engineers need to co-author the flow.

| RAG / LLM framework should I reach for?" Synthesizes the ecosystem sections of 7 landmark-project SOPs (LangGraph, LlamaIndex, DSPy, CrewAI, vLLM, Aider, Dify) a real project usually combines DSPy (compile) + LlamaIndex (retrieve) + LangGraph (orchestrate) + vLLM (serve), and you choose ONE per layer, not one to rule all. Use when starting any LLM/agent/RAG project, or whenever the "which framework?" question is asked. Deliberately neutral — unlike vendor docs and the LangChain-biased `framework-selection` on skill.sh, this skill has no horse in the race.

>- make the few reasoning decisions and a cheap model do the many mechanical executions (Aider architect+editor, DSPy optimizer-LM vs task-LM, vLLM speculative draft+target, LangGraph supervisor+worker are the same shape). Use when designing or cost-optimizing a pipeline that calls an LM many times, when deciding which steps need a strong reasoner vs a cheap executor, or when adding an escalation valve for when the cheap tier degrades. cheap model, strong model plus cheap model, LLM cost optimization.

| Operating-system distillation of LlamaIndex — the leading RAG / document-agent framework. Activate when the calling agent must build, debug, harden, or evaluate a Retrieval-Augmented Generation pipeline over unstructured/private data, decide between RAG primitives (Index types, retrievers, query engines, routers, agents), or pick LlamaIndex vs LangChain / Haystack / raw vector store for a coding task. Encodes the 5-layer mental model (Documents → Nodes → Indices → Retrievers → Query Engines / Response Synthesizers), the canonical RAG bootstrap SOP from baseline `VectorStoreIndex` through hybrid + reranker + eval-loop hardening, the official 13-failure-mode checklist, and 5 dilemma cases distilled from docs, GitHub issues, and 2025 production post-mortems.

| Enhancement-overlay SOP for adding sparse (BM25 / keyword) retrieval alongside dense (embedding) retrieval. Activate when a calling agent is building, reviewing, or debugging a retrieval pipeline whose corpus contains exact-match tokens — identifiers, error codes, SKUs, API/function names, proper nouns, citations, rare jargon — that pure dense embedding silently misses. Encodes the single decision share that depends on exact tokens is non-trivial**), the wiring of QueryFusionRetriever-style fusion (RRF vs alpha-weighted), and per-query-type alpha tuning. Frame the work as recovering lexical identity that dense pooling destroys, not as "add keyword search for completeness". Cross-links [[llamaindex]].

| Re-ingest-correctness SOP for production RAG. Activate when a calling agent builds, reviews, or debugs an ingestion pipeline that runs more than once over a changing corpus — scheduled re-index, incremental updates, CI re-ingest, or a "retrieval has duplicates / shows deleted docs" bug. Encodes the rule — insert/update/skip, so re-running over unchanged docs is a no-op** — plus the docstore + doc-hash upsert machinery (LlamaIndex `IngestionPipeline` + `DocstoreStrategy`), the delete-propagation problem, and cross-framework equivalents (LangChain `index()` + `RecordManager`, manual hash table). ENHANCE the re-ingest-correctness contract is not surfaced.

>- + compiled programs + model snapshot pins + retrieval config + eval-set version, versioned together so a deploy is reproducible and rollback is atomic. Activate when preparing to deploy an LLM app, when asking "what exactly is running in prod right now?", when a deploy must be reproducible months later, or when an incident needs a clean rollback. The core independently-mutable parts, not one weights file. Do NOT activate for one-off prompt edits with no deploy, for a single-component demo, or where a vendor owns the whole prompt lifecycle. For versioning ONE compiled prompt use [[agentsop-per-model-artifacts]]; for versioning, reproducible deploy, what is running in prod, rollback LLM app, model pinning, prompt registry, version prompts and config.

Cross-engine decision rubric for self-hosting or recommending an LLM serving stack. Picks among vLLM, SGLang, TensorRT-LLM, TGI, llama.cpp, Ollama, and MLX as a function of (hardware × workload × constraint), not "which is fastest". Activates whenever a coder-agent must choose, defend, or migrate a serving runtime.

>- Decision protocol for making side-effectful agent tools idempotent — so when an LLM tool call is retried (timeout, framework resume, user re-run, model duplicate emit), the second it'll call exactly once; the tool must promise the second call is safe. Framework-agnostic — applies to LangGraph node bodies that re-run on resume, MCP tools, OpenAI tool-calling duplicate email sent, charged twice, exactly-once, idempotency key, tool called twice, retry side effect, double-send, at-least-once delivery.

| Security-first SOP for multi-tenant RAG systems. Activate when a calling agent is building, reviewing, or debugging any retrieval pipeline whose vector store is shared across more than one user, organisation, workspace, customer, or permission scope. Encodes the single non-negotiable rule — **filter at the vector store query, never after retrieval / never after rerank** — together with the per-vendor query-time filter APIs (Pinecone namespaces + `$eq`/`$in`, Weaviate `multiTenancyConfig` + tenant handle, Qdrant `is_tenant` payload index + `Filter.must`, Chroma `where`, pgvector RLS), and the cross-framework adapters (LlamaIndex `MetadataFilters`, LangChain `filter=` dict). Frame the work as preventing CVE-2024-41892 / EchoLeak / Slack-AI-class cross-tenant leakage, not as "adding a filter for relevance".

>- Decomposed, multi-criteria metric design for LLM pipelines. The metric IS the model — change the metric and the optimizer changes behavior. Decompose by default; bool during LLM-as-judge, llm as judge, eval metric, evaluation score, scoring function, rubric, RAGAS, G-Eval, judge bias, verbosity bias, how to evaluate LLM output.

>- Enhancement-overlay (C5) for RAG over long documents — the chunk-paradox resolution. Activate when a single fixed chunk size cannot satisfy both retrieval precision (small large chunks dilute embedding relevance into "topic averages". Encodes the core flip — large for synthesis context — and the SOP to pick a base chunk size, choose a horizontal (sentence-window) vs vertical (auto-merging / parent-child) expansion strategy, and measure the lift. Cross-links [[llamaindex]] for the full RAG SOP; this overlay supplies the missing "chunk-paradox-resolution" recipe that the framework docs (HierarchicalNodeParser, SentenceWindow) only describe in fragments. Medium-frequency for chunking strategy, parent document retriever, sentence window, small-to-big retrieval, hierarchical chunking, optimal chunk size.

>- Lifecycle SOP for **per-model prompt artifacts** — the compiled prompts, instructions, few-shot demos, edit-format pins, and embedding-bound indices that change behavior when the underlying LM, dataset, or framework version changes. Activate when adopting compiled prompts (DSPy, GEPA, BootstrapFewShot output), when supporting multiple LMs in production, when a provider deprecates a model snapshot, or when a framework deprecates a config surface (LlamaIndex `ServiceContext` → `Settings`, Aider edit-format defaults). Do NOT activate for one-off raw prompt edits or for truly model-agnostic system prompts that have deprecation, prompt per model, prompt breaks on new model, version compiled prompts.

>- Enhancement-overlay SOP for query-type routing — sending a query to the right index / tool / engine *before* retrieving, not after. Activate when a calling agent owns a retrieval or answering surface that fronts more than one handler (a summary index, a vector index, a vs "find the clause about X" vs "how many orders shipped in Q3". Encodes the one non- negotiable insight — **one retriever cannot serve all query types; route first, retrieve second** — plus the three router families (LLM/selector, embedding/semantic, keyword/rule), confidence-threshold + fallback discipline, and the cross-framework mapping (LlamaIndex `RouterQueryEngine` / `SelectorPromptTemplate`, Dify Question Classifier node, LangGraph conditional edges). This is an ENHANCE overlay over the per-framework skills — cross-link `[[llamaindex]]`, `[[agentsop-dify]]`, `[[agentsop-langgraph]]` for the deep

>- Enhancement-overlay SOP for the reranker stage of a RAG pipeline — the "retrieve wide, rerank narrow" discipline. Activate when a calling agent owns a retrieval pipeline whose the context window is under pressure from too many marginal chunks. Encodes the one non- negotiable insight — a cheap bi-encoder retrieves *wide* for recall, then a more expensive cross-encoder (which reads query + document *together*) reranks *narrow* for precision; keep top-N=20-50, rerank to top-k=3-5. Covers when to add a reranker (and when not to), N-vs-k tuning, model choice (Cohere/Voyage API vs bge-reranker local vs SentenceTransformer cross-encoder), latency/cost budgeting, and the cross-framework mapping (LlamaIndex node postprocessors, LangChain ContextualCompressionRetriever, Cohere/Voyage rerank APIs, local cross-encoders). This is an ENHANCE overlay over the per- framework skills — cross-link `[[llamaindex]]` and `[[agentsop-hybrid-retrieval]]` for the

>- Project-kickoff rubric for the self-host vs managed-cloud decision — when is running your own inference engine / LLM platform worth the ops cost vs paying per-token for a managed API? Decide on two axes — VOLUME (a cost-crossover slider) and COMPLIANCE (a hard gate). Use at kickoff when choosing where to run inference, or when cost / data-residency pressure forces a re-evaluation.

Decision SOP for serving LLMs with vLLM. Covers PagedAttention mental model, quantization/parallelism/batching tradeoffs, OOM triage, and when NOT to use vLLM. Activates when a coder-agent is choosing or tuning an inference engine, debugging vLLM throughput/latency/OOM, or comparing vLLM against TGI/SGLang/TensorRT-LLM/llama.cpp.

获取 Hugging Face 上的模型、数据集和 Space 的统计信息

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

A specialized skill for embedding and extracting resilient watermarks in text by manipulating sentence lengths and using Fountain Codes. Use this when the user wants to add a hidden, robust watermark to text or verify an existing one.

能力蒸馏工作流——从用户批准的强模型访谈或真实任务轨迹中提取非显然的判断规则，形成可审计的 judgment packet，再交给 skill-audit 和 skill-creator 决定是否落成可加载 skill。当用户说“蒸馏这个模型的判断力”“把这次任务的关键决策固化下来”“模型窗口要关了，保留它在某类场景的判断”时使用。普通流程文档、直接写 SKILL.md、泛化最佳实践或未授权的会话日志扫描不使用本 skill。

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, update or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Deploy, evaluate, fine-tune, and manage Foundry agents end-to-end with azd: hosted agent scaffold/run/deploy, prompt agent create, batch eval, continuous eval, prompt optimizer, Agent Optimizer scaffold, agent.yaml, dataset curation from traces, model fine-tuning (SFT/DPO/RFT). USE FOR: azd ai agent, azd provision/deploy, deploy agent, hosted agent, create agent, add tool to agent, invoke agent, evaluate agent, continuous eval, continuous monitoring, agent CI/CD, optimize prompt, improve prompt, optimize agent instructions, agent optimizer, deploy model, Foundry project, RBAC, role assignment, permissions, quota, capacity, region, troubleshoot agent, deployment failure, AI Services, create Foundry resource, provision, knowledge index, customize deployment, onboard, availability, fine-tune, SFT, DPO, RFT, training-data, grader, distillation, fine-tuned model, large file upload. DO NOT USE FOR: Azure Functions, App Service, general Azure deploy (use azure-deploy), general Azure prep (use azure-prepare).

Advanced AI agent benchmark scenarios that push Vercel's cutting-edge platform features — Workflow DevKit, AI Gateway, MCP, Chat SDK, Queues, Flags, Sandbox, and multi-agent orchestration. Designed to stress-test skill injection for complex, multi-system builds.

End-to-end benchmark suite for vercel-plugin. Runs realistic projects through skill injection, launches dev servers, verifies everything works, analyzes conversation logs, and produces an improvement report for overnight self-improvement loops.

Run vercel-plugin eval scenarios in Vercel Sandboxes instead of local WezTerm panels. Provisions ephemeral microVMs with Claude Code + plugin pre-installed, runs benchmark prompts, extracts hook artifacts, and produces coverage reports.

Create and launch benchmark test projects to exercise vercel-plugin skill injection across realistic scenarios. Sets up isolated directories, installs the plugin, and spawns WezTerm panes running Claude Code with crafted prompts.

Vercel AI SDK expert guidance. Use when building AI-powered features — chat interfaces, text generation, structured output, tool calling, agents, MCP integration, streaming, embeddings, reranking, image generation, or working with any LLM provider.

Corrects outdated LLM knowledge about the Vercel platform and introduces new products. Injected at session start.

Patterns for invoking the Apify CLI (`apify`) from agents. Covers authentication, creating/running/pushing Actors, calling Actors in the cloud, and reading results from datasets and key-value stores.

Machine learning and LLM engineering judgment, distilled from a stronger model - invoke when DECIDING whether/how to use ML or an LLM for a task (prompt vs RAG vs fine-tune vs classical); working with training/eval data or labels; building or reviewing evals for models and LLM features; designing RAG, structured output, or agent pipelines; or diagnosing why a model/LLM feature underperforms. Method-selection ladder, data and leakage discipline, eval-as-spec rules, LLM-era craft, and a trap catalog.

AI agent and LLM system engineering reference covering single-agent dev (ReAct, tool calling, plan-execute), multi-agent coordination (swarm, role decomposition, file locking), LLM security (prompt injection, jailbreak defense, output filtering), RAG architecture (chunking, hybrid retrieval, rerank), and prompt engineering / evaluation (RAGAS, LLM-as-Judge). Use when building AI agents, designing RAG pipelines, orchestrating multi-agent workflows, hardening LLM apps, or writing prompts.

Use Databricks built-in AI Functions (ai_classify, ai_extract, ai_summarize, ai_mask, ai_translate, ai_fix_grammar, ai_gen, ai_analyze_sentiment, ai_similarity, ai_parse_document, ai_prep_search, ai_query, ai_forecast) to add AI capabilities directly to SQL and PySpark pipelines without managing model endpoints. Also covers document parsing and building custom RAG pipelines (parse → prep_search → index → query).

Train ML models on Databricks. Use for: classification/regression/deep-learning (XGBoost, scikit-learn, LightGBM, PyTorch) with Optuna, @prod/@challenger aliases, batch scoring (spark_udf for plain models, fe.score_batch for feature-store-backed), custom PyFunc, custom ResponsesAgent (LangGraph + UC Function/Vector Search); UC feature tables + FeatureLookup + point-in-time joins + Lakebase online store; declarative Feature Views (create_feature, DeltaTableSource, RollingWindow/SlidingWindow/TumblingWindow, materialize_features, streaming Kafka features). NOT for: endpoint ops (databricks-model-serving), MLflow evaluation (databricks-mlflow-evaluation).

Generate realistic synthetic data using Spark + Faker (strongly recommended). Supports serverless execution, multiple output formats (Parquet/JSON/CSV/Delta), and scales from thousands to millions of rows. For small datasets (<10K rows), can optionally generate locally and upload to volumes. Use when user mentions 'synthetic data', 'test data', 'generate data', 'demo dataset', 'Faker', or 'sample data'.

Build RAG / unstructured-document evaluation datasets and demo documents (e.g. for Knowledge Assistant) on Databricks: generate synthetic PDFs locally, upload to Unity Catalog volumes, and pair each document with test questions for retrieval evaluation.

Databricks Vector Search endpoints and indexes for RAG and semantic search; covers index types, search modes, end-to-end RAG patterns

Generates a comprehensive client health overview across all accounts. Reads CRM data, support tickets, usage metrics, billing, and engagement logs. Calculates health scores, trend direction, and RAG status per client. Outputs a sorted risk report with recommended actions.