Claude Skills 中文技能市场——为 Claude Code、Codex 与 Cursor 精选 Agent Skills

收录 90,000+ 开源 SKILL.md 工作流,提供中文解读、安全检测与高速下载,一行命令即可装到你的 AI 编程助手里。

加载中…

高分中文技能

ai-safety-auditor

eddiebe147's avatarfrom eddiebe147

对AI系统进行安全、偏见和负责任部署的审计。

data-training-manager

codatta's avatarfrom codatta

Manage AI training data, monitor content freshness, detect repetition, and update training samples for continuous learning. Use when managing training data, checking content quality, updating AI models, or preventing repetitive content.

edu-demo-evaluator

hanialshater's avatarfrom hanialshater

Evaluate educational demos using Chrome tools for E2E testing. Executes testcases from test_cases.json, captures screenshots, verifies learning outcomes.Scores QUALITY relative to benchmark. Uses real browser interaction viamcp__claude-in-chrome__* tools.

org-universal-validator

intent-solutions-io's avatarfrom intent-solutions-io

验证技能和插件,使用确定性证据包和严格模式门。用于审计变更或强制合规。通过“运行验证”或“审计验证器”触发。

eval-patterns

therealchrisrock's avatarfrom therealchrisrock

This skill provides common evaluation patterns and integration guidance. Use when:- Integrating eval-framework with other plugins- Designing evaluation workflows- Choosing between content vs behavior evaluation- Setting up project-local rubrics

nixtla-benchmark-reporter

intent-solutions-io's avatarfrom intent-solutions-io

Generate comprehensive markdown benchmark reports from forecast accuracy metrics with model comparisons, statistical analysis, and regression detection. Use when analyzing baseline performance, comparing forecast models, or validating model quality. Trigger with 'generate benchmark report', 'analyze forecast metrics', or 'create performance summary'.

prompt-iteration

darthmolen's avatarfrom darthmolen

Use when iteratively improving agent prompts through automated LLM-as-Judge evaluation. Runs eval→fix→commit loop with circuit breakers.

improve-skill

mauromedda's avatarfrom mauromedda

Analyze Claude Code session transcripts to improve existing skills or create new ones. Use when you want to review a past session to identify what worked, what didn't, and how to enhance skill documentation. Extracts session data and provides structured analysis prompts. Triggers on "improve skill", "analyze session", "review session", "skill improvement", "create skill from session", "skill not working", "skill missed", "skill didn't trigger", "enhance skill", "refine skill", "skill feedback", "session transcript", "what went wrong", "skill optimization", "better triggers".

evaluation-quality

majiayu000's avatarfrom majiayu000

仪器评估指标、质量分数和反馈循环

test-mcp-connector

majiayu000's avatarfrom majiayu000

仅当用户明确请求基于 MCP 的测试时才触发此技能:**必需的触发器(全部必须明确提及“MCP”):**-“使用 mcp 测试连接器”-“测试 mcp 连接器”-“使用 mcp 测试 [提供程序]”-“使用 mcp 测试 [提供程序]”-“运行 mcp 连接器测试”-“针对 [提供程序]”进行 mcp 测试“**不要触发:**- 通用“测试连接器”请求(使用 stackone 运行) / test_actions 相反) - 没有明确提及 MCP 的“测试 [provider]” - 定期验证或测试请求 - 任何未明确提及 MCP 的测试此技能使用 Claude Agent SDK 构建一个 REAL 代理,发送自然语言提示来评估操作描述是否对代理友好。它比常规测试更加密集,并且仅应在明确要求时使用。

rubric-design

therealchrisrock's avatarfrom therealchrisrock

This skill provides guidance on designing effective evaluation rubrics. Use when:- Creating criteria for content or code quality assessment- Defining weights and thresholds for evaluation- Designing check types (pattern-based vs custom)- Structuring rubrics for maintainability and reusability

marker-engine-rl

DYAI2025's avatarfrom DYAI2025

Vertieft den Marker-Engine-Skill um SFT/RL-Feinabstimmung mit LeanDeep 4.0; lädt Marker aus Supabase/ZIP und lernt eine Policy zur präzisen, kontextualisierten Marker-Anwendung bei strikter Bottom-up-Logik.

model-evaluator

eddiebe147's avatarfrom eddiebe147

Evaluate and compare ML model performance with rigorous testing methodologies

growth-learning

randysalars's avatarfrom randysalars

Analytics, feedback processing, and continuous improvement

nixtla-universal-validator

intent-solutions-io's avatarfrom intent-solutions-io

Validate Nixtla skills and plugins with deterministic evidence bundles and strict schema gates. Use when auditing changes or enforcing compliance. Trigger with 'run validation' or 'audit validators'.

pop-assessment-performance

jrc1883's avatarfrom jrc1883

使用具体指标评估PopKit的效率,包括上下文使用、令牌消耗和懒加载验证。

procedure-refiner

dakesan's avatarfrom dakesan

通过验证和重新生成周期迭代优化LA-Bench实验程序。在按ID改进生成的程序、根据LA-Bench标准验证程序质量以及管理程序检查器和程序生成器技能之间的优化循环时,应使用此技能。在要求优化、改进或验证输出JSONL文件中的LA-Bench程序时触发。

decision-critic

timmye's avatarfrom timmye

通过 python 脚本立即调用以对决策和推理进行压力测试。不要先分析 - 脚本会协调批评工作流程。

常见问题

什么是 Claude Skills(Agent Skills)?

Skills 是一种标准化的 SKILL.md 指令包——把「何时触发、如何执行、使用哪些工具」写成结构化工作流,装入 Claude Code、Codex、Cursor 等 AI 编程助手后,Agent 遇到对应任务会自动调用,效果远好于每次手写提示词。

装了 Skill 之后怎么使用?

安装后无需手动调用:直接在 Claude Code / Cursor 里描述你的任务(如「帮我做 PR 代码审查」),Agent 检测到匹配的技能后会自动加载对应 SKILL.md 并按其流程执行。也可以在对话中显式要求「使用 xx skill」。

国内如何高速下载安装?

本站在 Cloudflare R2 托管了全部技能 ZIP 离线包,无需直连 GitHub:在详情页点击「Download ZIP」,或用 `npx skills add <owner/repo>` 一键安装;也可以复制 curl 命令直接解压到 ~/.claude/skills 目录。