Dart Use Ffigen logo

Dart Use Ffigen

OrganizationPopular
flutter
dart-use-ffigen

Guide agents to use `package:ffigen` to automatically generate FFI bindings instead of writing them manually. Use this skill when a task involves writing new FFI bindings, extending C/Objective-C/Swift integrations, or replacing hand-crafted `dart:ffi` setups.

Overview

Publisherflutter
Repositoryagent-plugins
Skill namedart-use-ffigen
Stars
3K
Forks
182
Bundled files
Instructions only
LicenseBSD-3-Clause
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by flutter on GitHub. Read the source before you install it.

Installation

Install the Dart Use Ffigen AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/flutter/agent-plugins/tree/main/skills/dart-use-ffigen
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/flutter/agent-plugins.git /tmp/agent-plugins
mkdir -p .claude/skills
cp -r /tmp/agent-plugins/skills/dart-use-ffigen .claude/skills/dart-use-ffigen
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Dart Use Ffigen in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Dart Use Ffigen on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Dart Use Ffigen is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Generating FFI Bindings using package:ffigen

Contents

Introduction

Automate and standardize the generation of FFI bindings using package:ffigen (FfiGenerator). Writing FFI bindings by hand is error-prone, brittle, and highly discouraged.

Constraints

  • No Hand-Written FFI Bindings: If native headers (.h files) exist or are generated by a build step, never write manual DynamicLibrary.lookup, @Native external functions, or raw struct classes. Always use FfiGenerator to generate them.
  • Generator Location: The generator script should be located at tool/ffigen.dart within the target package root.
  • Header Location: If the native header files are third-party, they should be located in third_party/ within the target package (otherwise placing them in a src/ directory at the package root is also acceptable). If the headers are not in one of these standard locations, notify the user that it would be cleaner to move the header files to the standard location (e.g., third_party/).
  • Targeted Inclusion Filters: Avoid importing an entire native library unless specifically needed. Always apply precise inclusion lists using positive matches to minimize the size and cognitive load of the generated code (e.g., using Functions.includeSet or filtering matches in include closures).
  • Output Setup: If the generated FFI bindings interface with a third-party library (or reference third-party headers), the generated files must always be placed under lib/src/third_party/. The primary generated FFI bindings file must strictly use the .g.dart extension (e.g. sqlite3.g.dart).
  • Preamble & License Headers: Always supply a premium preamble in the Output class to specify the license. This must match the native third-party library's license, explicitly include the copyright header of the target native header file, and contain an automatic generation warning (e.g. // Generated by package:ffigen. Do not edit manually.).
  • No Unnecessary Commits of Stale Bindings: Ensure you run the generator script and check if the generated files have changed before finishing your task. Always verify the package by running dart analyze.
  • Record Usage and Tree Shaking: If the package is integrated into standard runtime execution or compiles native assets via native hooks:
    • Enable recorded usage on all functions by setting recordUse: (_) => true under Functions.
    • Specify the recordUseMapping target in Output (which must strictly be a .g.dart file under lib/src/third_party/, e.g. lib/src/third_party/sqlite3.record_use_mapping.g.dart) to register bindings for symbol tree shaking.

FFIgen Overview

To construct the programmatic generator, use the core configuration objects imported from package:ffigen/ffigen.dart:

1. FfiGenerator

The parent class that orchestrates the configuration, parsing, and code generation.

dart
FfiGenerator({
  Headers headers = const Headers(),
  Enums enums = Enums.excludeAll,
  Functions functions = Functions.excludeAll,
  Globals globals = Globals.excludeAll,
  Integers integers = const Integers(),
  Macros macros = Macros.excludeAll,
  Structs structs = Structs.excludeAll,
  Typedefs typedefs = Typedefs.excludeAll,
  Unions unions = Unions.excludeAll,
  UnnamedEnums unnamedEnums = UnnamedEnums.excludeAll,
  ObjectiveC? objectiveC,
  required Output output,
}).generate();

2. Headers

Configures Clang header parsing targets and compiler flags.

  • entryPoints: A list of target header Uri inputs.
  • include: A filter function bool Function(Uri header) that handles transitive header imports.
  • compilerOptions: Custom preprocessor/include compiler flags to pass directly to libclang.
  • ignoreSourceErrors: Set to true to silence errors occurring inside third-party headers during parsing.

3. Functions

Specifies which native C/C++ functions to expose in Dart.

  • include: A matcher function (e.g. (decl) => {'my_func'}.contains(decl.originalName) or Functions.includeSet({'my_func'})).
  • isLeaf: Declares functions as leaf functions ((decl) => true) if they do not call back into Dart or block thread execution.
  • recordUse: Enables metadata generation for native asset tree shaking (essential in dart-lang/native). Set to (_) => true.

4. Output

Configures target generated files.

  • dartFile: Target Uri where the primary FFI bindings will be written.
  • recordUseMapping: Target Uri for recorded usage metadata maps (crucial for linking-time tree shaking).
  • preamble: Text inserted at the top of the generated file (licensing, annotations).
  • format: Set to true to run the Dart formatter automatically.

Step-by-Step Workflow

Step 1: Check/Add Dependencies

Open the package's pubspec.yaml and verify the dev_dependencies contains ffigen. Use the Dart MCP server or look up the latest version on pub.dev (e.g., ^20.1.1).

You can add it automatically using the CLI:

bash
dart pub add dev:ffigen

Step 2: Formulate Paths Dynamically

Create a programmatic generator script under the package's tool/ directory (e.g., tool/ffigen.dart). Resolve paths relative to Platform.script to make sure it runs successfully from any working directory:

dart
final packageRoot = Platform.script.resolve('../');
final headerFile = packageRoot.resolve('third_party/library.h');
final targetBindings = packageRoot.resolve('lib/src/third_party/bindings.g.dart');

Step 3: Write the Script (tool/ffigen.dart)

Define void main() and run FfiGenerator with dynamic options (see complete example below).

Step 4: Run Code Generation

Execute the script from the terminal inside the target package folder:

bash
dart run tool/ffigen.dart

Step 5: Static Analysis

Verify that the generated bindings are correct and resolve any analysis issues. FFIgen automatically runs the Dart formatter on the output file (via format: true configuration), so manual formatting is not required.

  1. Run the static analyzer inside the target package:
    bash
    dart analyze
  2. Addressing Warnings/Lints: If dart analyze reports style or lint warnings inside the generated file, append the corresponding warning codes to the ignore_for_file: list in your generator script's preamble configuration (e.g., adding camel_case_types, non_constant_identifier_names, etc.). Do not modify the package's global rules.
  3. Addressing Compilation Errors: If dart analyze reports actual compiler or analysis errors (not warnings) inside the generated file, do not attempt to edit the generated file manually. Report these error details directly to the user so they can file an issue on the repository at github.com/dart-lang/native.

Concrete Example: Binding a C Library

Let's assume we are working with the SQLite package under pkgs/code_assets/example/sqlite, which embeds SQLite C library sources inside third_party/sqlite/ and accesses it via FFI.

The C Header File (third_party/sqlite/sqlite3.h)

c
// The author disclaims copyright to this source code.

#ifndef SQLITE3_H_
#define SQLITE3_H_

const char *sqlite3_libversion(void);

#endif // SQLITE3_H_

BEFORE: Manual FFI Binding (The Anti-Pattern)

A developer might attempt to handcraft this integration. It is fragile, blocks tree-shaking metadata, and is highly prone to ABI and structural mapping issues:

dart
// lib/src/sqlite3_manual.dart
import 'dart:ffi' as ffi;
import 'package:ffi/ffi.dart';

// Flaw 1: Hardcoded DynamicLibrary lookup blocks integration with modern native asset compilation.
final ffi.DynamicLibrary _dylib = ffi.DynamicLibrary.open('libsqlite3.so');

// Flaw 2: Manual function type matching requires writing redundant dynamic lookup boilerplate and lacks tree-shaking metadata.
typedef _sqlite3_libversion_C = ffi.Pointer<ffi.Char> Function();
typedef _sqlite3_libversion_Dart = ffi.Pointer<ffi.Char> Function();

final _sqlite3_libversion_Dart sqlite3LibVersion = _dylib
    .lookup<ffi.NativeFunction<_sqlite3_libversion_C>>('sqlite3_libversion')
    .asFunction();

AFTER: Generating via FFIgen (The Correct Pattern)

Create a programmatic script at tool/ffigen.dart:

dart
// Copyright (c) 2025, the Dart project authors.  Please see the AUTHORS file
// for details. All rights reserved. Use of this source code is governed by a
// BSD-style license that can be found in the LICENSE file.

import 'dart:io';
import 'package:ffigen/ffigen.dart';

void main() {
  // Resolve paths dynamically relative to Platform.script
  final packageRoot = Platform.script.resolve('../');
  final entryHeader = packageRoot.resolve('third_party/sqlite/sqlite3.h');
  final bindingsOutput = packageRoot.resolve('lib/src/third_party/sqlite3.g.dart');
  final treeShakeMapping = packageRoot.resolve('lib/src/third_party/sqlite3.record_use_mapping.g.dart');

  FfiGenerator(
    headers: Headers(
      entryPoints: [entryHeader],
    ),
    functions: Functions(
      include: (decl) => {'sqlite3_libversion'}.contains(decl.originalName),
      // Essential for package optimization and tree-shaking
      recordUse: (_) => true,
    ),
    output: Output(
      dartFile: bindingsOutput,
      recordUseMapping: treeShakeMapping,
      format: true,
      preamble: '''

// AUTO-GENERATED FILE - DO NOT MODIFY.
// Generated via ffigen.
// To regenerate: dart run tool/ffigen.dart

// ignore_for_file: type=lint, unused_import, unused_element, deprecated_member_use_from_same_package, experimental_member_use
''',
    ),
  ).generate();

  print('Successfully generated sqlite3 FFI bindings.');
}

Running the Generator script

Run this in the package root directory:

bash
dart run tool/ffigen.dart

This will automatically create:

  1. lib/src/third_party/sqlite3.g.dart
  2. lib/src/third_party/sqlite3.record_use_mapping.g.dart

Verification Checklist

Always perform the following verification before completing a binding generation task:

  1. Correct Setup: Verify the target generated files are inside lib/src/third_party/ (required for third-party licensed code) and the primary FFI bindings file strictly uses the .g.dart extension.
  2. Static Analysis: Run dart analyze and ensure there are zero compiler/analyzer errors or warnings in the package.
    • If static warnings are reported in the generated bindings, suppress them by adding ignore_for_file rules to the generator's preamble configuration (do not modify global package rules).
    • If actual compiler or analyzer errors are reported in the generated bindings, do not edit the generated file manually. Report the details to the user and direct them to file an issue at github.com/dart-lang/native.

Frequently asked questions

What does the Dart Use Ffigen AI skill do?

Guide agents to use `package:ffigen` to automatically generate FFI bindings instead of writing them manually. Use this skill when a task involves writing new FFI bindings, extending C/Objective-C/Swift integrations, or replacing hand-crafted `dart:ffi` setups.

Why use Dart Use Ffigen on TypingMind?

Because you install it once and use it with any model. Dart Use Ffigen is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Dart Use Ffigen in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/flutter/agent-plugins/tree/main/skills/dart-use-ffigen. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Dart Use Ffigen?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Dart Use Ffigen?

As many as you like. As long as a model supports skills, you can use Dart Use Ffigen with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Dart Use Ffigen AI skill free?

Yes. It is published on GitHub by flutter under the BSD-3-Clause license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇