turbo_promptable #

Object-Oriented Prompting framework for the turbo ecosystem. Define AI agent prompts, roles, workflows, and tools as type-safe Dart objects that serialize to JSON, YAML, Markdown, and XML.

Features #

• Type-safe workspace models: Role, Persona, Agent, Workflow, Step, Activity, Instruction, Input, Output, Goal, Issue, Context, Template, Tool, and more
• Spec models: Ability, Feature, Requirement, Scenario, Journey, Task, Module, Mockup, Prototype
• Spawnable abstraction (TSpawnable) with CLI tool and prompt-delivery enums for launching agents against Claude Code, Cursor, Windsurf, and others
• Cross-referencing abstracts (OfAbilities, OfFeatures, OfIssues, OfJourneys, OfMockups, OfModules, OfPrds, OfProjects, OfPrototypes, OfScenarios) for composing specs
• JSON serialization on every model via json_serializable; YAML, Markdown, and XML output inherited from turbo_serializable
• Structured metadata (TMetaData) and body/config control (TConfig) on every promptable

Installation #

dependencies:
  turbo_promptable: ^0.3.0

Usage #

import 'package:turbo_promptable/turbo_promptable.dart';

const instruction = Instruction(
  name: 'Code Quality',
  rules: ['No unused imports', 'All public API must have dartdoc'],
  principles: ['Clarity over cleverness'],
);

const workflow = Workflow(
  name: 'Review Workflow',
  steps: [
    Step(
      name: 'Analyse',
      input: Input(
        name: 'Source Code',
        request: 'Analyse the provided source code for quality issues.',
      ),
      instructions: [instruction],
      output: Output(name: 'Analysis Report'),
    ),
  ],
);

const role = Role(
  name: 'Code Reviewer',
  expertise: 'Static analysis and code quality',
  instructions: [instruction],
  workflows: [workflow],
);

final agent = Agent.fromRole(
  role,
  identity: 'A meticulous reviewer focused on maintainability.',
);

print(agent.toJson());

Core Concepts #

TPromptable #

Every workspace model extends TPromptable, which itself extends TSerializable from turbo_serializable. Models have:

• name — required identifier
• metaData — optional TMetaData for frontmatter (description, tags, etc.)
• config — optional TConfig for body rendering and inheritance

Roles, Personas, and Agents #

• Role — a capability bundle with expertise, activities, checklists, instructions, templates, tools, and workflows
• Persona — a Role augmented with an identity string
• Agent — a spawnable Persona with CLI tool and command targeting, constructible via Agent.fromRole(...) or Agent.fromPersona(...)

Workflows and Steps #

A Workflow contains an ordered list of Steps. Each Step has an Input, optional Instructions, and an Output.

Specs #

Spec models (Ability, Feature, Requirement, Scenario, Journey, Task, Module, Mockup, Prototype) describe intended behaviour and deliverables. They cross-reference each other via the Of* abstracts exported from workspace/abstracts/.

Spawnable #

TSpawnable (used by Role, Persona, Agent) carries cliTool (TCliTool), command, and promptDelivery (TPromptDelivery) for orchestrating agent launches across tools like Claude Code, Cursor, Windsurf, and custom CLIs.

License #

MIT