Skip to content
Design Token Authority

Commands

The help text of every command, as dta <command> --help prints it. Exit codes: Exit codes.

$ dta --help
Usage: dta [options] [command]


Design Token Authority — sync Figma variables to code, and back.


Options:
  -V, --version               output the version number
  -h, --help                  display help for command


Commands:
  pull [options]              Export variables from Figma to local token JSON
                              files
  push [options]              Push local token JSON files to Figma
  brief [options]             A short, prompt-ready brief of each brand for AI
                              design tools: colours with their roles, type,
                              spacing, radii, breakpoints, rules
  prune [options]             Delete variables from the Figma file whose tokens
                              are no longer in the token files (push never
                              deletes)
  build [options]             Generate CSS, JS, and other outputs from local
                              token files
  analyze [options]           Analyze Figma file structure and infer layer roles
  graph [options]             Build and visualize the token alias dependency
                              graph
  clean [options]             Remove all token files and build output
  lint [options]              Validate token files against linting rules
  diff [options]              Compare the tokens with a base (directory, git
                              ref, or merge base with main)
  token                       Look up tokens: what a token resolves to, which
                              tokens have a value, which exist
  skill                       Install the DTA skill that teaches coding agents
                              the token workflow
  audit [options] [paths...]  Find hard-coded colours, lengths and fonts in code
                              that equal a token, and name the token to use
  context [options]           Write DESIGN.md for coding agents (this project's
                              brands, tokens to use and rules) and point
                              AGENTS.md to it
  init [options]              Create a dta.config.ts and set up dta diff; in a
                              project with a config, set up or update dta diff
  help [command]              display help for command

dta pull

Section titled “dta pull”

Export variables from Figma to local token JSON files.

$ dta pull --help
Usage: dta pull [options]


Export variables from Figma to local token JSON files


Options:
  -o, --output <dir>   Output directory (overrides config tokens.dir)
  --from-file <path>   Import from a plugin-exported JSON file instead of the
                       Figma API
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta push

Section titled “dta push”

Push local token JSON files to Figma.

$ dta push --help
Usage: dta push [options]


Push local token JSON files to Figma


Options:
  --dry-run            Show what would change without modifying Figma
  --format <format>    Diff output format: console, markdown, json (default:
                       "console")
  --skip-lint          Skip lint check before pushing
  -y, --yes            Skip confirmation prompt (use in CI/automation)
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta brief

Section titled “dta brief”

A short, prompt-ready brief of each brand for AI design tools: colours with their roles, type, spacing, radii, breakpoints, rules.

$ dta brief --help
Usage: dta brief [options]


A short, prompt-ready brief of each brand for AI design tools: colours with
their roles, type, spacing, radii, breakpoints, rules


Options:
  --brand <name>       Only this brand (default: one brief per brand)
  --format <format>    markdown (to paste into a prompt) or json (to pipe into a
                       tool) (default: "markdown")
  --copy               Also put the brief on the clipboard
  -c, --config <path>  Path to config file
  -h, --help           display help for command

dta prune

Section titled “dta prune”

Delete variables from the Figma file whose tokens are no longer in the token files (push never deletes).

$ dta prune --help
Usage: dta prune [options]


Delete variables from the Figma file whose tokens are no longer in the token
files (push never deletes)


Options:
  --dry-run            List the orphaned variables without deleting anything
  --force              Delete without the typed confirmation (CI, with explicit
                       opt-in)
  --allow-bulk         Also delete in a collection that would lose more than
                       half of its variables (left alone by default)
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta build

Section titled “dta build”

Generate CSS, JS, and other outputs from local token files.

$ dta build --help
Usage: dta build [options]


Generate CSS, JS, and other outputs from local token files


Options:
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta analyze

Section titled “dta analyze”

Analyze Figma file structure and infer layer roles.

$ dta analyze --help
Usage: dta analyze [options]


Analyze Figma file structure and infer layer roles


Options:
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta graph

Section titled “dta graph”

Build and visualize the token alias dependency graph.

$ dta graph --help
Usage: dta graph [options]


Build and visualize the token alias dependency graph


Options:
  -f, --format <format>  Output format: console, dot, markdown, html (default:
                         "console")
  -o, --output <path>    Write output to file instead of stdout
  -c, --config <path>    Path to config file
  -v, --verbose          Enable verbose logging
  -h, --help             display help for command

dta clean

Section titled “dta clean”

Remove all token files and build output.

$ dta clean --help
Usage: dta clean [options]


Remove all token files and build output


Options:
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta lint

Section titled “dta lint”

Validate token files against linting rules.

$ dta lint --help
Usage: dta lint [options]


Validate token files against linting rules


Options:
  --fix                Auto-fix violations where possible
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta diff

Section titled “dta diff”

Compare the tokens with a base (directory, git ref, or merge base with main).

$ dta diff --help
Usage: dta diff [options]


Compare the tokens with a base (directory, git ref, or merge base with main)


Options:
  --base <ref|dir>       Base to compare with (default: merge base of HEAD and
                         the default branch)
  --tokens-only          Only compare token values, no screenshots
  --list                 List the stories dta diff would render, then exit
  --check                Check the setup without rendering (config, browser, git
                         history, base, stories); one fix per problem
  --report <dir>         Folder for screenshots and results (default: .dta/diff)
  --publish <command>    Publish hook: runs on the report folder and prints its
                         public URL (overrides diff.publish)
  --from-report          Publish the report folder of an earlier run instead of
                         rendering again
  -f, --format <format>  Output format: console, markdown, json (default:
                         "console")
  -o, --out <file>       Write the markdown or JSON report to a file
  --fail-on-change       Exit with code 1 when something needs a look:
                         unexplained story changes above the threshold,
                         explained ones to review, new or removed stories (token
                         changes without stories: any change)
  -c, --config <path>    Path to config file
  -v, --verbose          Enable verbose logging
  -h, --help             display help for command

dta token get

Section titled “dta token get”

Show a token: type, collection, value and alias chain per brand or mode, and its names in CSS and Tailwind.

$ dta token get --help
Usage: dta token get [options] <name>


Show a token: type, collection, value and alias chain per brand or mode, and its
names in CSS and Tailwind


Arguments:
  name                             Token name in dot form, e.g. color.action.bg


Options:
  -m, --mode <collection=mode...>  Read in these modes only, e.g. --mode
                                   Brand=BrandB (repeatable)
  -f, --format <format>            Output format: console, json (default:
                                   "console")
  -c, --config <path>              Path to config file
  -v, --verbose                    Enable verbose logging
  -h, --help                       display help for command

dta token find

Section titled “dta token find”

Find the tokens that resolve to a value: at every screen width first, aliases (semantic and component tokens) before primitives.

$ dta token find --help
Usage: dta token find [options] <value>


Find the tokens that resolve to a value: at every screen width first, aliases
(semantic and component tokens) before primitives


Arguments:
  value                  A colour (#2f55d4, rgb(47 85 212)), number, string or
                         boolean


Options:
  -t, --type <type>      Only tokens of this type: color, number, string,
                         boolean
  --near                 Colours: also match colours within about 2% per channel
  -f, --format <format>  Output format: console, json (default: "console")
  -c, --config <path>    Path to config file
  -v, --verbose          Enable verbose logging
  -h, --help             display help for command

dta token list

Section titled “dta token list”

List tokens with type and collection, sorted by name.

$ dta token list --help
Usage: dta token list [options]


List tokens with type and collection, sorted by name


Options:
  --collection <name>    Only tokens of this collection
  -t, --type <type>      Only tokens of this type: color, number, string,
                         boolean
  --match <glob>         Only names matching the glob: * within a segment, **
                         across segments
  -f, --format <format>  Output format: console, json (default: "console")
  -c, --config <path>    Path to config file
  -v, --verbose          Enable verbose logging
  -h, --help             display help for command

dta skill add

Section titled “dta skill add”

Write the DTA skill for coding agents (SKILL.md: look tokens up, lint, build, diff, never push).

$ dta skill add --help
Usage: dta skill add [options]


Write the DTA skill for coding agents (SKILL.md: look tokens up, lint, build,
diff, never push)


Options:
  --dir <path>  Folder to write SKILL.md to (other agents than Claude Code)
                (default: ".claude/skills/dta")
  --check       Exit 1 when the installed skill is missing or out of date
  -h, --help    display help for command

dta audit

Section titled “dta audit”

Find hard-coded colours, lengths and fonts in code that equal a token, and name the token to use.

$ dta audit --help
Usage: dta audit [options] [paths...]


Find hard-coded colours, lengths and fonts in code that equal a token, and name
the token to use


Arguments:
  paths                  Files and folders to scan (default: audit.paths, else
                         src)


Options:
  --near                 Also report colours within about 2% per channel of a
                         token
  -f, --format <format>  Output format: console, markdown, json (default:
                         "console")
  --update-baseline      Record the current findings as accepted; later runs
                         report only new ones
  --baseline <file>      Baseline file (default: audit.baseline, else
                         .dta/audit-baseline.json)
  -c, --config <path>    Path to config file
  -v, --verbose          Enable verbose logging
  -h, --help             display help for command

dta context

Section titled “dta context”

Write DESIGN.md for coding agents (this project’s brands, tokens to use and rules) and point AGENTS.md to it.

$ dta context --help
Usage: dta context [options]


Write DESIGN.md for coding agents (this project's brands, tokens to use and
rules) and point AGENTS.md to it


Options:
  -o, --out <file>     Context file to write (default: context.out in the
                       config, else DESIGN.md)
  --check              Exit 1 when the context file or the AGENTS.md block is
                       missing or out of date
  -c, --config <path>  Path to config file
  -v, --verbose        Enable verbose logging
  -h, --help           display help for command

dta init

Section titled “dta init”

Create a dta.config.ts and set up dta diff; in a project with a config, set up or update dta diff.

$ dta init --help
Usage: dta init [options]


Create a dta.config.ts and set up dta diff; in a project with a config, set up
or update dta diff


Options:
  --diff         Only the dta diff step (what dta init does in a project with a
                 config): detect, write the diff section, the GitHub workflow
                 and playwright-core; run again to update
  -y, --yes      No questions (CI and coding agents): without a config, a
                 dta.config.ts for exported token files with every output
                 format; with --diff or a config, the detected or default
                 answers
  -v, --verbose  Enable verbose logging
  -h, --help     display help for command