migrate-filemaker

A Claude Code skill that guides full migrations from FileMaker Pro to modern open-source stacks.

Give it a FileMaker Database Design Report (DDR) XML export and it will parse every component, walk you through requirements discovery, recommend a tech stack, and produce a complete migration plan with SQL schema, API design, and UI spec.

Requirements

• Claude Code CLI
• Python 3.6+ — used by the DDR XML parser (scripts/parse_ddr.py)

Installation

Copy or symlink into your Claude Code skills directory:

# Personal skills (available in all projects)
cp -r migrate-filemaker ~/.claude/skills/migrate-filemaker

# Or project-local skills
cp -r migrate-filemaker .claude/skills/migrate-filemaker

Usage

From Claude Code, provide a path to your DDR export:

/migrate-filemaker /path/to/ddr-export/

The input can be:

• A single FMPReport type="Report" XML file
• A directory containing one or more Report XMLs (multi-file DDR)

Multi-file solutions (e.g. UI file + data file) are auto-detected. All files are parsed and merged into unified specs.

What it does

The skill runs in four phases, each producing checkpoint files so work can resume across sessions:

Phase 1: Parse & Understand

Runs parse_ddr.py on the DDR XML and extracts 9 JSON spec files:

File	Contents
00_topology.json	File inventory, cross-file references
01_tables.json	Tables with fields, calculated fields, summaries, globals
02_table_occurrences.json	Table occurrences (graph nodes)
03_relationships.json	Relationships with join predicates
04_layouts.json	Layouts with fields, portals, buttons
05_scripts.json	Scripts with parsed steps and parameters
06_value_lists.json	Value lists (custom and field-based)
07_security.json	Accounts and privilege sets
08_custom_functions.json	Custom functions with calculations

Then analyzes the specs and produces an application summary with complexity scoring, data model overview, feature map, and red flags.

Phase 2: Discovery

Interactive requirements gathering — asks about goals, users, UI preferences, technical constraints, and integrations. Adapts questions based on earlier answers.

Phase 3: Recommend

Recommends a tech stack (database, backend, frontend, auth, deployment), architecture pattern, migration strategy, and feature priority order.

Phase 4: Plan

Generates the full migration plan:

Output	Description
03_migration_plan.md	Phased plan with effort estimates and risk register
04_database_schema.sql	Production-ready SQL DDL
05_api_design.md	RESTful API endpoints grouped by domain
06_ui_spec.md	Frontend spec with page inventory, component mapping, form specs

Project structure

migrate-filemaker/
  SKILL.md                  # Skill definition (instructions for Claude)
  scripts/
    parse_ddr.py            # DDR XML parser (Python)
  reference/
    ddr-xml-reference.md    # DDR XML structure documentation
    filemaker-concepts.md   # FileMaker-to-modern concept mapping
    schema-translation-guide.md
    script-translation-patterns.md
    tech-stack-decision-matrix.md
  templates/
    00_app_summary.md       # Output templates for each phase
    01_discovery_answers.md
    02_recommendations.md
    03_migration_plan.md
    04_database_schema.sql
    05_api_design.md
    06_ui_spec.md

DDR parser standalone usage

The parser can also be run directly outside of Claude Code:

# Parse a single file
python3 scripts/parse_ddr.py /path/to/report.xml

# Parse a directory of DDR XMLs
python3 scripts/parse_ddr.py /path/to/ddr-export/

# Specify output directory
python3 scripts/parse_ddr.py /path/to/ddr-export/ /path/to/output/

If no output directory is given, specs are written to a specs/ folder next to the input.

License

MIT