文本提及网络访问
8 次命中此源文件行包含 URL 或网络客户端。应检查目标地址、发送的数据和信任边界;静态文本无法证明实际发生了网络活动。 该规则在收集文本中命中 8 处;此处列出 3 个代表位置。命中文本可能只是文档,分析器没有执行它。
技能研究档案
Designs and implements production-grade ML pipeline infrastructure: configures experiment tracking with MLflow or Weights & Biases, creates Kubeflow or Airflow DAGs for training orchestration, builds feature store schemas with Feast, deploys model registries, and automates retraining and validation workflows. Use when building ML pipelines, orchestrating training workflows, automating model lifecycle, implementing feature stores, managing experiment tracking systems, setting up DVC for data versioning, tuning hyperparameters, or configuring MLOps tooling like Kubeflow, Airflow, MLflow, or Prefect.
Designs and implements production-grade ML pipeline infrastructure: configures experiment tracking with MLflow or Weights & Biases, creates Kubeflow or Airflow DAGs for training orchestration, builds feature store schemas with Feast, deploys model registries, and automates retraining and validation workflows. Use when building ML pipelines, orchestrating training workflows, automating model lifecycle, implementing feature stores, managing experiment tracking systems, setting up DVC for data versioning, tuning hyperparameters, or configuring MLOps tooling like Kubeflow, Airflow, MLflow, or Prefect.
Skill Atlas editorial review · editorial-preview-1.0.0
用于构建生产级机器学习流水线的 Skill,涵盖实验跟踪、编排、特征存储以及部署前的评估关卡。
对所收集的 SKILL.md 及辅助文本/代码进行确定性静态分析。机器辅助候选和发现等待人工编辑复核;不评估动态行为或效果。
文件清单来自记录提交中的采集源码。路径与大小属于结构观察,不代表质量评价。
skills/ml-pipeline/references/experiment-tracking.md引用21.4 KBskills/ml-pipeline/references/feature-engineering.md引用17.4 KBskills/ml-pipeline/references/model-validation.md引用29.8 KBskills/ml-pipeline/references/pipeline-orchestration.md引用23 KBskills/ml-pipeline/references/training-pipelines.md引用22 KBskills/ml-pipeline/SKILL.md指令7.2 KB{
"name": "ml-pipeline",
"description": "Designs and implements production-grade ML pipeline infrastructure: configures experiment tracking with MLflow or Weights & Biases, creates Kubeflow or Airflow DAGs for training orchestration, builds feature store schemas with Feast, deploys model registries, and automates retraining and validation workflows. Use when building ML pipelines, orchestrating training workflows, automating model lifecycle, implementing feature stores, managing experiment tracking systems, setting up DVC for data versioning, tuning hyperparameters, or configuring MLOps tooling like Kubeflow, Airflow, MLflow, or Prefect.",
"license": "MIT",
"metadata": {
"author": "https://github.com/Jeffallan",
"company": "https://synergetic.solutions",
"version": "1.1.0",
"domain": "data-ml",
"triggers": "ML pipeline, MLflow, Kubeflow, feature engineering, model training, experiment tracking, feature store, hyperparameter tuning, pipeline orchestration, model registry, training workflow, MLOps, model deployment, data pipeline, model versioning",
"role": "expert",
"scope": "implementation",
"output-format": "code",
"related-skills": "devops-engineer, kubernetes-specialist, cloud-architect, python-pro"
}
}此工作流根据源码材料重建表面顺序。除非明确记录受控评估,否则它属于暂定解读,而非真实运行轨迹。
梳理数据流,划分阶段并定义组件间接口。
校验失败时停止并报告,而不是继续训练。
提供流水线定义、特征代码、带跟踪的训练脚本、评估阈值和部署配置。
静态发现用于提示安装前需要检查的位置,不能证明一项技能安全、恶意或有效。
收集的文本中有 164 项确定性词法发现需要复核。发现可能仅位于文档中,不证明已执行或可被利用;仅含二进制的文件未扫描。
此源文件行包含 URL 或网络客户端。应检查目标地址、发送的数据和信任边界;静态文本无法证明实际发生了网络活动。 该规则在收集文本中命中 8 处;此处列出 3 个代表位置。命中文本可能只是文档,分析器没有执行它。
此行包含可能的文件写入或重定向操作。应检查目标路径和覆盖行为;静态审查没有写入任何文件。 该规则在收集文本中命中 154 处;此处列出 3 个代表位置。命中文本可能只是文档,分析器没有执行它。
此行出现 shell、动态求值或进程执行原语。应检查参数和输入控制;源代码中出现这些语法并不证明命令已运行。 该规则在收集文本中命中 2 处;此处列出 2 个代表位置。命中文本可能只是文档,分析器没有执行它。
运行生成的流水线代码前,检查凭据、数据路径和 pickle 模型产物的存储与加载位置。
发现、激活与资源层具有不同含义。只有使用指定官方方法并记录模型的运行时,才显示精确数值。
在管线使用明确模型执行官方精确计数之前,包括文件清单在内的所有 token 值均为 null。
未使用要求的官方方法记录 token 测量。本档案不会用字符数、文件大小或估计值替代。
该技能仅具备以上标明的证据。不声称任何任务成功率、token 效率、延迟或质量优势。
git clone https://github.com/Jeffallan/claude-skills.git
cd claude-skills
# 将该 Skill 目录复制到以下任一位置:
# ~/.claude/skills/ml-pipeline
# .claude/skills/ml-pipeline让 Claude Code 为训练脚本添加 MLflow 跟踪,确认其记录超参数和指标并先校验数据。
Using the ml-pipeline skill, build a Kubeflow training pipeline for a classifier with data validation, MLflow tracking, an evaluation gate with explicit thresholds, and a rollback strategy.
Read the ml-pipeline skill at https://github.com/Jeffallan/claude-skills/blob/67c1961b180dc92cf08983344627109e983e8f51/skills/ml-pipeline/SKILL.md and its reference files. Explain its workflow, any commands it tells the agent to run, and the main risks before I decide whether to install it.
研究结论适合已在使用或计划使用 MLOps 工具的团队的、注重可复现性的指南。
档案是对公开工件的限时阅读,而非背书。请通过来源链接检查确切的上游材料。
67c1961b180dc92cf08983344627109e983e8f51仅对提供的文件进行确定性词法解析、YAML frontmatter 校验、文件清单、来源描述的工作流程提取和启发式暂定分类。由机器生成;无 LLM、执行、动态测试或人工编辑批准。