Agent Skills › open-metadata/OpenMetadata › connector-standards

connector-standards

GitHub

加载OpenMetadata连接器开发标准，涵盖架构、代码风格、测试及数据源规范。用于指导或审查连接器开发，确保模式一致性和最佳实践遵循。

skills/connector-standards/SKILL.md open-metadata/OpenMetadata

触发场景

用户请求加载标准开始连接器开发前进行连接器代码审查时

安装

npx skills add open-metadata/OpenMetadata --skill connector-standards -g -y

更多选项

不安装直接使用

npx skills use open-metadata/OpenMetadata@connector-standards

指定 Agent (Claude Code)

npx skills add open-metadata/OpenMetadata --skill connector-standards -a claude-code -g -y

安装 repo 全部 skill

npx skills add open-metadata/OpenMetadata --all -g -y

预览 repo 内 skill

npx skills add open-metadata/OpenMetadata --list

SKILL.md

Frontmatter

{
    "name": "connector-standards",
    "description": "Load all OpenMetadata connector development standards into context. Use before building or reviewing connectors to ensure consistent patterns.",
    "allowed-tools": [
        "Read",
        "Glob"
    ],
    "argument-hint": "[optional: specific standard name like 'testing' or 'database']",
    "user-invocable": true
}

Load OpenMetadata Connector Standards

When to Activate

When a user asks to "load standards", "show connector standards", or before starting any connector development or review work.

Behavior

Load All Standards

If no specific standard is requested, load all standards in this order:

${CLAUDE_SKILL_DIR}/standards/main.md — Architecture overview
${CLAUDE_SKILL_DIR}/standards/patterns.md — Error handling, logging, pagination
${CLAUDE_SKILL_DIR}/standards/code_style.md — Python and JSON Schema conventions
${CLAUDE_SKILL_DIR}/standards/schema.md — Connection schema patterns
${CLAUDE_SKILL_DIR}/standards/connection.md — Connection class patterns
${CLAUDE_SKILL_DIR}/standards/service_spec.md — ServiceSpec registration
${CLAUDE_SKILL_DIR}/standards/testing.md — Unit and integration test patterns
${CLAUDE_SKILL_DIR}/standards/registration.md — How to register a connector
${CLAUDE_SKILL_DIR}/standards/performance.md — Performance best practices
${CLAUDE_SKILL_DIR}/standards/memory.md — Memory management and OOM prevention
${CLAUDE_SKILL_DIR}/standards/lineage.md — Lineage extraction methods
${CLAUDE_SKILL_DIR}/standards/sql.md — SQLAlchemy patterns and URL building

Then read all source-type standards:

${CLAUDE_SKILL_DIR}/standards/source_types/*.md

Load Specific Standard

If a specific standard or service type is requested:

Request	File to Load
"testing"	`standards/testing.md`
"patterns"	`standards/patterns.md`
"schema"	`standards/schema.md`
"lineage"	`standards/lineage.md`
"sql"	`standards/sql.md`
"memory"	`standards/memory.md`
"database"	`standards/source_types/database.md`
"sql databases"	`standards/source_types/sql_databases.md`
"data warehouses"	`standards/source_types/data_warehouses.md`
"nosql"	`standards/source_types/nosql_databases.md`
"dashboard"	`standards/source_types/dashboard.md`
"pipeline"	`standards/source_types/pipeline.md`
"messaging"	`standards/source_types/messaging.md`
"mlmodel"	`standards/source_types/mlmodel.md`
"storage"	`standards/source_types/storage.md`
"search"	`standards/source_types/search.md`
"api"	`standards/source_types/api.md`
etc.	`standards/source_types/{name}.md`

After Loading

Confirm to the user which standards were loaded and summarize the key points. Example:

Loaded 12 core standards + 11 source-type standards. Key points:

Schema-first: one JSON Schema → Python, Java, TypeScript, UI forms

Use BaseConnection for SQLAlchemy, get_connection()/test_connection() for others

Use pytest with plain assert, no unittest.TestCase

Always include copyright header, use ingestion_logger()

Lineage via query logs (database), SQL parsing (dashboard), or task metadata (pipeline)