MasterGo Skill

Retrieve and analyze design data from MasterGo files using self-contained Python scripts.

Prerequisites

Environment variable required:

export MASTERGO_TOKEN="mg_your_token_here"

Get token: MasterGo Settings > Security > Personal Access Token

Requirements: Team Edition account, files in Team Projects (not Drafts).

CRITICAL: Secure Token Handling

NEVER do this:

• echo $MASTERGO_TOKEN - exposes secret in output
• echo "Token: $MASTERGO_TOKEN" - exposes secret
• printenv MASTERGO_TOKEN - exposes secret
• Any command that prints the token value

Secure verification method:

test -n "$MASTERGO_TOKEN" && echo "Token is set" || echo "Token is NOT set"

This only prints whether the token exists, never its value.

Script Location

Scripts are in scripts/ relative to this SKILL.md file. Use absolute path when executing:

{this_skill_directory}/scripts/mastergo_*.py

Workflow

Step 1: Analyze DSL Structure

Always analyze first to understand the page structure:

python scripts/mastergo_analyze.py "https://mastergo.com/goto/xxx"

Output shows:

• Node tree (type, name, size)
• Text contents
• Component doc links
• Navigation targets

Step 2: Get Full DSL (if needed)

For detailed DSL data:

python scripts/mastergo_get_dsl.py "https://mastergo.com/goto/xxx"

Output: JSON with { dsl, componentDocumentLinks, rules }

Step 3: Fetch Component Docs

If componentDocumentLinks is non-empty, fetch relevant docs:

python scripts/mastergo_get_dsl.py URL | python scripts/mastergo_fetch_docs.py --from-dsl

Or fetch individually:

python scripts/mastergo_fetch_docs.py "https://example.com/button.mdx"

Scripts Reference

Script	Purpose	Output
mastergo_analyze.py	Structure summary	Human-readable tree to stdout
mastergo_get_dsl.py	Full DSL data	JSON to stdout
mastergo_fetch_docs.py	Component docs	Doc content to stdout
mastergo_utils.py	Utility functions	Import as module

DSL Key Concepts

• token fields: Design tokens for colors, shadows, fonts (convert to CSS variables when generating code)
• componentInfo: Component metadata including documentation links
• interactive fields: User interactions including page navigation targets
• rules array: Guidelines returned with DSL response

CRITICAL: Do NOT Pollute User Projects

FORBIDDEN actions:

• Creating .temp_dsl.json or any DSL cache file in user project
• Creating analyze_dsl.py or any analysis script in user project
• Creating any temporary files in user project directory
• Writing any skill-related files outside of skill directory

ALLOWED:

• Running scripts and capturing stdout output
• Using output directly in memory
• Reading user's existing project files for context

All script output goes to stdout only. Use the output directly; do not save intermediate files to user's project.

References

• references/dsl-types.md - Complete DSL type definitions
• references/dsl-structure.md - Key fields and patterns
• references/multi-page-workflow.md - Multi-page workflow