Agent skill

Jazz Schema Design

by garden-co in garden-co/classic-jazz

Design and implement collaborative data schemas using the Jazz framework.

MITAuto-check passedDatabases

Install Jazz Schema Design

skills CLI
$ npx skills add garden-co/classic-jazz --skill jazz-schema-design -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install garden-co/classic-jazz jazz-schema-design --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/garden-co/classic-jazz.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.cursor/skills/jazz-schema-design .claude/skills/jazz-schema-design && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
jazz-schema-design
GitHub stars
2.5k
Token cost
~3.2k tokens
SKILL.md length
1,077 words
Files
1
Skills in repo
8
Repo updated
First seen
Licence
MIT

At a glance

Design and implement collaborative data schemas using the Jazz framework.

  • Works in 4 steps: none - Cannot read content → reader - Can read content → writer - Can update content (overwrite… → …
  • Working with Jazz apps to define data structures using CoValues
  • SKILL.md covers When to Use This Skill, Do NOT Use This Skill For, Core Concepts and Schema Definition Basics, plus 10 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Jazz Schema Design is an agent skill from garden-co/classic-jazz. Design and implement collaborative data schemas using the Jazz framework. Use this skill when building or working with Jazz apps to define data structures using CoValues. This skill focuses exclusively on schema definition and data modeling logic.

Its SKILL.md is about 3.2k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Databases, covering Database schema design. The repository describes itself as: A new kind of database that's distributed across your frontend, containers, serverless functions and its own storage cloud. The licence is MIT.

When your agent uses it

  • Working with Jazz apps to define data structures using CoValues
  • Tasks that involve Database schema design

Example prompts

  • “/jazz-schema-design”

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

  1. none - Cannot read content
  2. reader - Can read content
  3. writer - Can update content (overwrite values, modify lists)
  4. admin - Can grant/revoke permissions plus all writer capabilities

What it can do on your machine

Read from SKILL.md and the folder at commit 4f90501. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md (its code samples are typescript).

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • jazz.tools
    • github.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Jazz Schema Design loads about 3.2k tokens when it runs. Until then it costs about 67 tokens; SKILL.md has 1,077 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~67
When it runs · the whole SKILL.md, loaded when a task matches
~3.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from garden-co/classic-jazz at commit 4f90501, republished under its MIT licence (© garden-co). 1,077 words, ~3,162 tokens.

Download SKILL.mdSave it as .claude/skills/jazz-schema-design/SKILL.md (or your agent's skills folder).
name
jazz-schema-design
description
Design and implement collaborative data schemas using the Jazz framework. Use this skill when building or working with Jazz apps to define data structures using CoValues. This skill focuses exclusively on schema definition and data modeling logic.

Jazz Data Modelling and Schema Design

When to Use This Skill

  • Designing data structures for Jazz applications
  • Defining CoValue schemas and relationships
  • Configuring permissions at the schema level
  • Planning schema evolution and migrations
  • Choosing between scalar and collaborative types
  • Modeling relationships between data entities

Do NOT Use This Skill For

  • Writing tests for Jazz applications (use the jazz-testing skill)
  • General framework integration questions (use the jazz-ui-development skill)
  • Permissions topics outside the schema. Use the jazz-permissions-security skill.

Key Heuristic for Agents: If the user is asking about how to structure their data model, define relationships, configure default permissions, or evolve schemas, use this skill.

Core Concepts

Jazz models data as an explicitly linked collaborative graph, not traditional tables or collections. Data types are defined as schemas, with CoValues serving as the fundamental building blocks.

Schema Definition Basics

Basic Structure
ts
const Author = co.map({
  name: z.string()
});

const Post = co.map({
  title: z.string(),
  content: co.richText(),
});

Key Libraries:

  • z.* - Zod schemas for primitive types (e.g., z.string())
  • co.* - Jazz collaborative data types (e.g., co.richText(), co.map())

Permissions Model

Permissions are integral to the data model, not an afterthought. Each CoValue has an ownership group with hierarchical permissions.

Permission Levels (cumulative)
  1. none - Cannot read content
  2. reader - Can read content
  3. writer - Can update content (overwrite values, modify lists)
  4. admin - Can grant/revoke permissions plus all writer capabilities
Critical Permission Rules
  • Only the creator is admin by default - no superuser concept
  • Cannot change CoValue ownership group - must modify group membership instead
  • Different permissions require separate containers - use distinct CoMaps/CoLists
  • Nested CoValues inherit permissions from parent by default
  • Define default permissions at schema level during design
Defining Permissions at Schema Level

Use withPermissions() to set automatic permissions when creating CoValues:

ts
const Dog = co.map({
  name: z.string(),
}).withPermissions({
  onInlineCreate: "sameAsContainer",
});

const Person = co.map({
  pet: Dog,
}).withPermissions({
  default: () => Group.create().makePublic(),
});

// Person CoValues are public, Dog shares owner with Person
const person = Person.create({
  pet: { name: "Rex" }
});
Permission Configuration Options

default Defines group when calling .create() without explicit owner.

onInlineCreate Controls behavior when CoValue is created inline (NOT applied to .create() calls):

  • "extendsContainer" (default) - New group includes container owner as member, inheriting permissions
  • "sameAsContainer" - Reuse container's owner (performance optimization—see below for concerns and considerations)
  • "newGroup" - New group with active account as admin
  • { extendsContainer: "reader" } - Like "extendsContainer" but override container owner's role
  • Custom callback - Create and configure new group as needed

onCreate Callback runs on every CoValue creation (both .create() and inline). Use to configure owner.

Global Permission Defaults

Set defaults for all schemas using setDefaultSchemaPermissions:

ts
import { setDefaultSchemaPermissions } from "jazz-tools";

setDefaultSchemaPermissions({
  onInlineCreate: "sameAsContainer",  // Performance optimization
});

USE EXTREME CAUTION: If you use sameAsContainer, you MUST be aware that the child and parent groups are one and the same. Any changes to the child group will affect the parent group, and vice versa. This can lead to unexpected behavior if not handled carefully, where changing permissions on a child group inadvertently results in permissions being granted to the parent group and any other siblings created with the same parent. As ownership cannot be changed, you MUST NOT USE sameAsContainer if you AT ANY TIME IN FUTURE may wish to change permissions granularly on the child group.

CoValue Types

TypeScript TypeCoValueUse Case
objectCoMapStruct-like objects with predefined keys
Record<string, T>CoRecordDict-like objects with arbitrary string keys
T[]CoListOrdered lists
T[] (append-only)CoFeedSession-based append-only lists
stringCoPlainText/CoRichTextCollaborative text editing
Blob | FileFileStreamFile storage
Blob | File (image)ImageDefinitionImage storage
number[] | Float32ArrayCoVectorEmbeddings/vector data
T | U (discriminated)DiscriminatedUnionMixed-type lists

Use the special types co.account() and co.profile() for user accounts and profiles.

Choosing Scalar vs Collaborative Types

Scalar Types (Zod: z.*)

Use when:

  • Full replacement updates expected
  • No collaborative editing needed
  • Single writer scenario
  • Raw performance critical

Examples:

ts
const myCoValue = co.map({
  title: z.string()        // Replace entire title, no collaboration needed
  coords: z.object({
    lat: z.number(),
    lon: z.number()
  })   // Replace entire object, no collaboration needed
});

Note: not all Zod types are available in Jazz. Be sure to always import { z } from 'jazz-tools';, and validate whether the type exists on the export. DO NOT import from zod.

Collaborative Types (Jazz: co.*)

Use when:

  • Multiple users edit simultaneously
  • Surgical/granular edits needed
  • Full edit history tracking valuable
  • Collaborative features required

Examples:

ts
const myCoVal = co.map({
  content: co.richText()   // Multiple editors
  items: co.list(Item)     // Add/remove individual items
  config: co.map({
    settingA: z.boolean(),
    settingB: z.number()
  })       // Update specific keys
});

Trade-off: CoValues track full edit history. Slightly slower for single-writer full-replacement scenarios, but benefits almost always outweigh costs.

Relationship Modeling

One-Directional Reference
ts
const Post = co.map({
  title: z.string(),
  author: Author  // One-way reference (like foreign key)
});

Jazz stores referenced ID. Use resolve queries to control reference traversal depth.

Show full SKILL.md (426 more words)Show less
Recursive/Forward References

Use getters to defer schema evaluation:

ts
const Author = co.map({
  name: z.string(),
  get posts() {
    return co.list(Post);  // Deferred evaluation
  }
});

const Post = co.map({
  title: z.string(),
  author: Author
});

Important: Jazz doesn't create inferred inverse relationships. Explicitly add both sides for bidirectional traversal.

Inverse Relationships (Two-Way)

One-to-One:

ts
const Author = co.map({
  name: z.string(),
  get post() {
    return Post;
  }
});

const Post = co.map({
  title: z.string(),
  author: Author
});

One-to-Many:

ts
const Author = co.map({
  name: z.string(),
  get posts() {
    return co.list(Post);
  }
});

const Post = co.map({
  author: Author
});

Many-to-Many:

Use co.list() at both ends. Jazz doesn't maintain consistency - manage in application code.

Set-Like Collections (Unique Constraint)

CoLists allow duplicates. For uniqueness, use CoRecord keyed on ID:

ts
const Author = co.map({
  name: z.string(),
  posts: co.record(z.string(), Post)
});

// Usage
author.posts.$jazz.set(newPost.$jazz.id, newPost);

Note: CoRecords always use string keys. Validate IDs at application level.

Data Discovery Pattern

CoValues are only addressable by unique ID. Discovery without ID requires reference traversal.

Standard pattern:

  • Attach 'root' CoValue to user account (entry point to the data graph)
  • For global 'roots': hardcode ID or use environment variable
  • Build graph from root via references

Schema Evolution

Each CoValue copy is authoritative. Users may be on different schema versions simultaneously.

Best Practices
  1. Add version field to schema
  2. Only add fields, never remove
  3. Never change existing field types
  4. Make new fields optional (backward compatible)
  5. Use withMigration() carefully - runs on every load
Migration Example
ts
const Post = co.map({
  version: z.number().optional(),
  title: z.string(),
  content: co.richText(),
  tags: z.array(z.string()).optional()  // New optional field
}).withMigration((post) => {
  // Exit early if already migrated
  if (post.version === 2) return;
  
  // Perform migration
  if (!post.$jazz.has('tags')) {
    post.$jazz.set('tags', []);
  }
  post.$jazz.set('version', 2);
});

Migration warnings:

  • Runs every time CoValue loads
  • Exit early to avoid unnecessary work
  • Poor migrations can significantly slow app

Design Checklist

  • Identify which data needs collaborative editing
  • Map permissions requirements to CoValue containers
  • Choose scalar vs collaborative types appropriately
  • Define explicit relationships (both directions if needed)
  • Plan root CoValue attachment strategy
  • Add version field for future evolution
  • Set default permissions at schema level
  • Handle recursive references with getters
  • Consider migration strategy for schema changes
  • Ensure an initial migration exists for the user account to ensure the profile and root are initialized
  • Ensure there are no TS or linting errors

Common Patterns

Blog with Authors and Posts
ts
const Author = co.map({
  name: z.string(),
  bio: co.richText(),
  get posts() {
    return co.list(Post);
  }
});

const Post = co.map({
  title: z.string(),
  content: co.richText(),
  author: Author,
  publishedAt: z.date().optional()
});
Collaborative Task List
ts
const Task = co.map({
  title: z.string(),
  description: co.richText(),
  completed: z.boolean(),
  assignees: co.list(User)
});

const Project = co.map({
  name: z.string(),
  tasks: co.list(Task)
});
User Profile with Settings
ts
const UserRoot = co.map({
  theme: z.literal(['light', 'dark']),
});

const UserProfile = co.profile({
  name: z.string(),
  bio: co.richText(),
  posts: co.record(z.string(), Post)
});

const UserAccount = co.account({
  profile: UserProfile,
  root: UserRoot
}).withMigration((account, creationProps) => {
  if (!account.has('root')) {
    const root = UserRoot.create({
      theme: 'light'
    });
    account.$jazz.set('root', root)
  }
  if (!account.has('profile')) {
    const profile = UserProfile.create({
      name: creationProps?.name ?? 'Anonymous User',
      bio: '',
      posts: {}
    })
  }
});

Anti-Patterns to Avoid

❌ Don't mix permissions in single CoValue - use separate containers ❌ Don't rely on inferred inverse relationships - explicitly define both sides ❌ Don't change field types in schema updates ❌ Don't write expensive migrations - they run on every load ❌ Don't use CoValues everywhere without considering trade-offs ❌ Don't forget to make new fields optional for backward compatibility

Quick Reference

Loading with relationships: Use resolve queries to control depth when traversing references.

Permission changes: Admin/manager modifies group membership, not CoValue ownership.

Unique IDs: Each CoValue has unique ID - only way to directly address without traversal.

Nested CoValues: Inherit permissions from parent when created inline.

References

Load these on demand, based on need:

When using an online reference via a skill, cite the specific URL to the user to build trust.

© garden-co, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .cursor/skills/jazz-schema-design of garden-co/classic-jazz.

Open the folder on GitHubat commit 4f90501

Compare with similar skills

Jazz Schema Design next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Jazz Schema Design compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Jazz Schema Design this skillgarden-co/classic-jazz2.5k—~3.2kAutomated safety check: PassMIT
SQL Optimization Patternsynulihao/AgentSkillOS61811 repos~3.3kAutomated safety check: PassNone
Datamodellmnimbalyst/nimbalyst1.9k—~713Automated safety check: PassMIT
Add Mpk Taskmirage-project/mirage2.5k—~4.5kAutomated safety check: PassApache-2.0
B200 Flash Attention4 Plannermirage-project/mirage2.5k—~1.9kAutomated safety check: PassApache-2.0
Experiment Auditwanshuiyin/Auto-claude-code-research-in-sleep17k1 repos~2.7kAutomated safety check: NotesMIT

Similar skills

  • SQL Optimization Patterns

    ynulihao/AgentSkillOS

    Master SQL query optimization, indexing strategies, and EXPLAIN analysis to dramatically improve database performance and eliminate slow queries.

    618 GitHub starsUsed in 11 repos~3.3k tokens
    DatabasesAuto-check passed
  • Datamodellm

    nimbalyst/nimbalyst

    Create visual data models for database schemas using Nimbalyst's DataModelLM editor.

    1.9k GitHub stars~713 tokensUpdated yesterday
    DatabasesAuto-check passed
  • Add Mpk Task

    mirage-project/mirage

    Step-by-step guide for adding a new task implementation to Mirage Persistent Kernel (MPK).

    2.5k GitHub stars~4.5k tokensUpdated 3 days ago
    DatabasesAuto-check passed
  • B200 Flash Attention4 Planner

    mirage-project/mirage

    A skill your agent uses when the user wants to design or extend a FlashAttention-style forward kernel on B200/Blackwell, involving the two MMAs QKᵀ and PV, online softmax, S/P/O in TMEM, warp roles…

    2.5k GitHub stars~1.9k tokensUpdated 3 days ago
    DatabasesAuto-check passed
  • Experiment Audit

    wanshuiyin/Auto-claude-code-research-in-sleep

    Audit experiment integrity before claiming results. An agent skill from wanshuiyin/Auto-claude-code-research-in-sleep.

    17k GitHub starsUsed in 1 repo~2.7k tokens
    DatabasesAuto-check: notes
  • Sqlite Schema Design

    fastrepl/anarlog

    Design or review schemas for crates/cloudsync using SQLite Sync constraints, not generic SQLite advice.

    9.5k GitHub stars~1.9k tokensUpdated today
    DatabasesAuto-check passed

More from garden-co/classic-jazz

All 8 skills in this repo
  • Jazz Performance

    garden-co/classic-jazz

    A skill your agent uses when optimizing Jazz applications for speed, responsiveness, and scalability.

    2.5k GitHub stars~1.9k tokensUpdated 1 mo ago
    Auto-check passed
  • Jazz Testing

    garden-co/classic-jazz

    A skill your agent uses when you need to write, review, or debug automated tests for applications built on the Jazz framework.

    2.5k GitHub stars~1.6k tokensUpdated 1 mo ago
    Auto-check passed
  • Jazz UI Development

    garden-co/classic-jazz

    A skill your agent uses when building, debugging, or optimizing Jazz applications.

    2.5k GitHub stars~1.7k tokensUpdated 1 mo ago
    Auto-check passed
  • Benchmarking

    garden-co/classic-jazz

    A skill your agent uses when writing or running performance benchmarks for Jazz packages.

    2.5k GitHub stars~3.2k tokensUpdated 1 mo ago
    Auto-check passed
  • Jazz Permissions Security

    garden-co/classic-jazz

    A skill your agent uses when designing data schemas, implementing sharing workflows, or auditing access control in Jazz applications.

    2.5k GitHub stars~2.9k tokensUpdated 1 mo ago
    Auto-check passed
  • Changeset

    garden-co/classic-jazz

    Generate changeset files for versioning and changelog management in this monorepo.

    2.5k GitHub stars~538 tokensUpdated 1 mo ago
    Auto-check passed

Categories

Questions about Jazz Schema Design

What does Jazz Schema Design do?

Design and implement collaborative data schemas using the Jazz framework. Jazz Schema Design is an agent skill from garden-co/classic-jazz. Design and implement collaborative data schemas using the Jazz framework.

When should I use Jazz Schema Design?

Jazz Schema Design fits situations like: working with Jazz apps to define data structures using CoValues; tasks that involve Database schema design.

How do I install Jazz Schema Design in Claude Code?

Run `npx skills add garden-co/classic-jazz --skill jazz-schema-design -a claude-code`. Or copy the skill folder (.cursor/skills/jazz-schema-design in garden-co/classic-jazz) into .claude/skills/jazz-schema-design in your project. Claude Code loads it when a task matches its description.

How do I install Jazz Schema Design in Codex?

Run `npx skills add garden-co/classic-jazz --skill jazz-schema-design -a codex`. Or copy the skill folder (.cursor/skills/jazz-schema-design in garden-co/classic-jazz) into .agents/skills/jazz-schema-design in your project. Codex loads it when a task matches its description.

Can I use Jazz Schema Design in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add garden-co/classic-jazz --skill jazz-schema-design -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/jazz-schema-design, .gemini/skills/jazz-schema-design, .github/skills/jazz-schema-design and .opencode/skills/jazz-schema-design in your project.

What does Jazz Schema Design need to run?

SKILL.md names no scripts, command-line tools or credentials: Jazz Schema Design is instructions for the agent only.

Does Jazz Schema Design access the network?

SKILL.md names 2 domains. As links in the text: jazz.tools and github.com. This is read from the text; nothing was executed.

Is Jazz Schema Design safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Jazz Schema Design use?

Jazz Schema Design is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Jazz Schema Design use?

About 3.2k tokens (SKILL.md is roughly 13k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Jazz Schema Design?

Skills that share tags, products or a category with Jazz Schema Design: SQL Optimization Patterns (ynulihao/AgentSkillOS, 618 stars), Datamodellm (nimbalyst/nimbalyst, 1.9k stars), Add Mpk Task (mirage-project/mirage, 2.5k stars) and B200 Flash Attention4 Planner (mirage-project/mirage, 2.5k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Jazz Schema Design?

garden-co (a GitHub organization) maintains it in garden-co/classic-jazz, which has 2,534 GitHub stars. The repository holds 8 skills in this directory. The repository was last updated on September 7, 2026.

Source: garden-co/classic-jazz on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.