Token导航 LogoToken导航

Agent Skill · 待分类

schema-manager(模式管理器)

schema-manager 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

schema-manager 是待分类类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在待分类类的 11,209 个 Skill 里排第 6,387。

总安装 461周安装 19GitHub Stars 2下载 150核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:schema-manager(模式管理器)
来源仓库:https://github.com/lytics/agent-skills
仓库路径:skills/schema-manager
安装命令:
npx skills add https://github.com/lytics/agent-skills --skill schema-manager
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/lytics/agent-skills --skill schema-manager

这个 Skill 是做什么的

这个 Skill 的分类还没最终确定,用法以来源仓库的说明为准。

用于处理 GitHub 仓库、Issue 和代码协作信息,适合围绕项目状态进行整理。可协助分析 Pull Request、跟踪 Issue 进展或梳理代码变更历史。使用时需结合来源仓库 README 了解具体能力边界,避免误判协作事项。安装前应确认权限范围,注意是否会触发文件读写或命令执行操作。

schema-manager 属于待分类类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/schema-manager

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/lytics/agent-skills --skill schema-manager;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

schema-manager 怎么安装?

在终端执行 npx skills add https://github.com/lytics/agent-skills --skill schema-manager。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

schema-manager 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

schema-manager 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

schema-manager 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

schema-manager 用的人多吗?

公开数据是累计安装 461 次,近一周安装 19 次,下载 150 次,GitHub 2 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

schema-manager 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理待分类类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

待分类类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
infsh-cli

infsh CLI · Codex、Claude

4,374,576382权限未标注未明确未收录
ai-video-generation

AI 视频生成 · Codex、Claude

4,313,376382调用外部服务未明确有提醒
lark-openapi-explorer

云雀 OpenAPI 浏览器 · Codex、Claude

2,064,1209,011需要联网未明确有提醒
lark-workflow-standup-report

Lark 工作流程站立报告 · Codex、Claude

2,041,6328,978需要联网未明确通过
use-my-browser

使用我的浏览器 · Codex、Claude

1,848,49140操作浏览器未明确未通过
schema-manager

模式管理器 · 本页

4612需要联网未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Schema Manager

Purpose

Browse and modify the profile schema -- fields, mappings, identity configuration, and field rankings. Use this when users want to understand their data model, add new fields, or modify field mappings.

Environment

Requires authenticated API access. See ../references/auth.md for credential resolution.

API Endpoints

Schema Overview

# List all schemas (tables)
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Get specific schema
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

Fields

# List all fields
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/field" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Get field names only
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/fieldnames" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Get specific field
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/field/${FIELD_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Create field
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/field" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "Field": "field_name",
    "Type": "string",
    "ShortDesc": "Brief description",
    "MergeOp": "latest",
    "IsIdentifier": false,
    "IsPII": false
  }'

# Update field -- requires the full field definition, not just changed fields.
# At minimum, include Field, Type, and the fields you're changing.
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/field/${FIELD_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "Field": "field_name",
    "Type": "string",
    "ShortDesc": "Updated description",
    "MergeOp": "latest",
    "IsIdentifier": true,
    "IsPII": false
  }'

# Delete field
curl -s -X DELETE "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/field/${FIELD_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

Mappings

A mapping connects a source stream field to a target schema field via an LQL expression. A field without at least one mapping is inert -- it will never receive data.

# List mappings
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/mapping" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Get specific mapping
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/mapping/${MAPPING_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Create mapping -- all three fields (field, stream, expr) are required
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/mapping" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "field": "email",
    "stream": "default",
    "expr": "email(email_address)"
  }'

# Create mapping with guard condition
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/mapping" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "field": "hashed_email",
    "stream": "click_stream",
    "expr": "hash.sha256(email(email))",
    "guard_expr": "email != '\'\''"
  }'

# Update mapping
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/mapping/${MAPPING_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "field": "email",
    "stream": "new_stream",
    "expr": "email(raw_email)"
  }'

# Delete mapping
curl -s -X DELETE "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/mapping/${MAPPING_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

Mapping payload fields:

FieldRequiredDescription
fieldYESTarget schema field name -- must reference an existing field
streamYESSource stream name (e.g., default, salesforce_contacts)
exprYESLQL expression to extract/transform the value (e.g., set(raw_field), email(email_address))
guard_exprNoOptional condition controlling when this mapping applies (do not prefix with "IF")

Common LQL expressions for mappings:

  • Simple pass-through: raw_field_name
  • Set/array: set(raw_field)
  • Email normalization: email(email_address)
  • Hashing: hash.sha256(email(email))
  • URL parsing: url(page_url)
  • Type casting: todate(timestamp_field)

Schema Patches

Schema patches group related changes into a named changeset that can be reviewed and applied together -- like a git branch for schema changes. Accounts with enable_schema_patches: true must use patches for all schema changes. See "Determining the Write Workflow" in the Behavior section.

Patch lifecycle: Create patch -> Add changes -> Review diff -> Apply

# List all patches for a table
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Create a new patch -- `tag` (kebab-case identifier) and `description` are both required.
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "tag": "add-purchase-fields",
    "description": "Add purchase tracking fields and mappings for Shopify integration"
  }'

# Get a specific patch (includes diffs against live schema)
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Update patch metadata
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{"description": "Updated description"}'

# Delete a patch (discard all changes)
curl -s -X DELETE "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Apply a patch (publishes a new schema version with all changes)
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}/apply" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

Add fields to a patch:

Patch endpoint payload shape is lowercase. Use id (the field name), type, shortdesc, mergeop, is_identifier, is_pii. The capitalized shape (Field, Type, ShortDesc,...) shown in the non-patch direct-write examples below is NOT accepted here -- patch endpoints will return Attribute 'Field' is required for Field if you send capitalized keys.
# Add or update a field in the patch
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}/field" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "id": "purchase_total",
    "type": "number",
    "shortdesc": "Total purchase amount",
    "mergeop": "sum",
    "is_identifier": false,
    "is_pii": false
  }'

# Update an existing field in the patch
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}/field/${FIELD_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{ ... full field definition ... }'

# Get a field from the patch (shows diff against live)
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}/field/${FIELD_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

Add mappings to a patch:

# Add a mapping in the patch
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}/mapping" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "field": "purchase_total",
    "stream": "shopify_orders",
    "expr": "total_price"
  }'

# Update an existing mapping in the patch
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}/mapping/${MAPPING_ID}" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{ ... mapping definition ... }'

Add rank changes to a patch:

curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/${TABLE}/${PATCH_ID}/rank" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{ ... rank config ... }'

Edit statuses within a patch:

StatusMeaning
newDoes not exist in live schema -- will be created
modifiedExists in live schema -- properties have been changed
deletedMarked for removal when patch is applied
unmodifiedIncluded for context, no changes from live

Identity Configuration

curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/idconfig" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

Field Rankings

# Get field rankings
curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/rank" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

# Update field rankings
curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/rank" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{ ... ranking config ... }'

Publish Schema Changes

Publish requires a tag and description. Both are mandatory.

curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/publish" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "tag": "short-kebab-tag",
    "description": "Human-readable description of what changed"
  }'

The tag should be a short kebab-case identifier (e.g., add-phone-identifier). The description should explain the changes for audit purposes.

Generate LQL from Schema

curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/lql" \
  -H "Authorization: ${LYTICS_API_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{ ... schema config ... }'

Field Types

See ../references/field-types.md for the complete type reference including:

  • Scalar: string, bool, int, number, date
  • Complex: []string, map[string]int, map[string]string, etc.
  • Special: geolocation, embedding, membership

Behavior

For Read Operations (list, get, browse)

Execute immediately. Present fields in a readable table format showing:

  • Field name, type, description, merge operation, identity/PII flags

Determining the Write Workflow

Some accounts require schema patches (account setting enable_schema_patches: true). Others use the direct publish workflow. You must determine which mode the account uses before making any schema changes.

Check by attempting to list patches:

curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/patch/user" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

If this returns successfully (even an empty list), the account uses patches. If it returns an error or 404, use the direct publish workflow.

Write Workflow A: Schema Patches (when enabled)

When the account has schema patches enabled, all schema changes must go through patches. Direct field/mapping edits will not work.

Complete workflow for adding a new field with mapping:

  1. Create a patch with a descriptive name
  2. Add the field to the patch
  3. Add a mapping for the field to the patch -- a field without a mapping is inert and will never receive data. Always ask: "Which stream should populate this field, and what is the source field name?"
  4. Review the patch diff (GET the patch to see all changes vs live schema)
  5. Confirm with the user (show the diff summary)
  6. Apply the patch (this publishes a new schema version)

If the user doesn't know the stream name, list available streams:

curl -s "${LYTICS_API_URL:-https://api.lytics.io}/v2/stream/names" \
  -H "Authorization: ${LYTICS_API_TOKEN}"

Write Workflow B: Direct Publish (when patches are not enabled)

When the account does not use schema patches, edits go into a shared unpublished draft and must be published manually. Schema changes are staged until published -- they do not take effect until you publish.

When creating a new field, always prompt for a mapping -- a field without a mapping is inert. Then publish both together.

Every write operation MUST follow this sequence:

  1. Confirm: Use the confirmation-gate pattern. Show what will change and get user approval.
  2. Execute: Make the field/mapping change via the API.
  3. Publish: Immediately publish the changes. A schema write is NOT complete until publish succeeds. curl -s -X POST "${LYTICS_API_URL:-https://api.lytics.io}/v2/schema/${TABLE}/publish" \ -H "Authorization: ${LYTICS_API_TOKEN}" \ -H "Content-Type: application/json" \ -d '{"tag": "descriptive-tag", "description": "What changed and why"}'
  4. Verify: Confirm publish succeeded. If it fails, report the error -- the changes are still staged and need to be published before they take effect.

Never consider a schema write operation complete without publishing. If multiple changes are being made in sequence, you may batch them and publish once at the end, but always publish before reporting success.

Error Handling

  • Field already exists (409): Show existing field details, ask if user wants to update instead
  • Invalid type: List valid types from ../references/field-types.md
  • Publish failures: Show error details, suggest checking field validity

Dependencies

  • Uses: ../references/auth.md, ../references/api-client.md, ../references/confirmation-gate.md
  • References: ../references/field-types.md

平台分布

Codex

32.79%
按下载量换算49

Claude

31.05%
按下载量换算47

Cursor

18.07%
按下载量换算27

Gemini CLI

9.65%
按下载量换算14

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills