turbo_promptable 0.4.0 copy "turbo_promptable: ^0.4.0" to clipboard
turbo_promptable: ^0.4.0 copied to clipboard

Object-Oriented Prompting framework for defining AI agent prompts, roles, workflows, and tools as type-safe Dart objects.

Changelog #

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.4.0 - 2026-04-17 #

Added #

  • Tool command model ToolCommand for declaring individual operations that a Tool exposes (CLI subcommands, REST endpoints, MCP functions, etc.)
  • Tool parameter models ToolParameter and ToolParameterOption for describing command inputs, enumerated options, defaults, and required flags
  • Tool ability model ToolAbility for describing the capabilities surfaced by a tool
  • Mcp tool model for declaring Model Context Protocol servers alongside the existing Api, Cli, and Script tool kinds
  • TMetaData is now re-exported from meta/t_promptable.dart so downstream models that depend on it no longer need a separate import

Changed #

  • BREAKING: Removed the Agent root model. Use Persona extended with TSpawnable fields (cliTool, command, promptDelivery) to represent a launchable agent.
  • BREAKING: Dropped the metaData constructor parameter from leaf context/root models that do not need it (Actor, Concept, Documentation, Project, Stakeholder, Subject, Checklist, Context, and related models). Pass metadata only on the models that still declare the parameter.
  • BREAKING: Removed barrel exports for internal/auxiliary members: core/constants/tp_keys.dart, core/models/t_config.dart, core/models/t_embed_type.dart, core/models/t_md_section.dart, core/models/t_render_type.dart, core/typedefs/t_body_builder_def.dart, workspace/models/checklists/acceptance_criteria.dart, workspace/models/checklists/constraints.dart, workspace/models/checklists/non_goals.dart, and workspace/models/meta/t_tag.dart. Consumers that relied on those names should import the deep paths directly or model their own equivalents.
  • TPromptable.mdFactory now derives its body from toJson() (minus the metadata key) so frontmatter/body separation is consistent across every subclass.
  • Bumped minimum turbo_serializable constraint to ^0.5.0 (pulls in the ts_map_extension rename and the wider barrel exposure from that release).

Fixed #

  • Removed unused and unnecessary imports flagged by the analyzer in workspace/models/context/collection.dart, workspace/models/root/activity.dart, and workspace/models/root/persona.dart.

0.3.0 - 2026-04-15 #

Changed #

  • BREAKING: Remove Dart source rendering from workspace entities (per-entity emission methods, helper mixins/classes under lib/core/helpers/, and the dedicated rendering test file). JSON, YAML, Markdown, and XML serialization remain.
  • Document public members on Ability so pub.dev documentation scoring stays at or above the 20% API threshold.

0.2.0 - 2026-04-13 #

Added #

  • TSpawnable base class for promptable models that can be spawned as CLI processes
  • TCliTool enum for supported CLI tools (Claude Code, Cursor, Windsurf, etc.)
  • TPromptDelivery enum for prompt delivery methods (system, user, stdin, file)
  • Spawn module with TFile, TFolder, and TSpawnConfig models
  • New root models: Activity, Agent, Checklist, Context, Goal, Input, Instruction, Issue, Output, Persona, PromptField, Role, Spec, Template, Tool, Workflow
  • New spec models: Mockup, Module, Prototype, Task with cross-referencing via Of* interfaces
  • TDartRenderHelper mixin for rendering models as Dart constructor calls
  • TDartStringHelper for Dart-safe string escaping
  • Workspace abstracts: OfAbilities, OfFeatures, OfIssues, OfJourneys, OfMockups, OfModules, OfPrds, OfProjects, OfPrototypes, OfScenarios
  • Project context model
  • Step workflow model with ordering and substep support

Changed #

  • BREAKING: Removed TTask root model — replaced by the richer Task spec model
  • BREAKING: Expanded Role from a simple model to a full spawnable with activities, checklists, instructions, templates, tools, and workflows
  • BREAKING: Persona and Agent now extend TSpawnable with CLI tool and command support
  • Enriched Activity with persona, goal, input, output, context, and workflow fields
  • Enriched Input with format, constraints, and validations fields
  • Enriched Output with format, constraints, validations, and deliverables fields
  • Enriched Instruction with scope, priority, applicability, and examples fields
  • Enriched spec models (Ability, Feature, Requirement, Scenario) with cross-referencing IDs

0.1.0 - 2026-04-10 #

Added #

  • Workspace model system with 8 domain groups: checklists, context, instructions, memories, meta, root, specs, tools, and workflows
  • Checklist models: TAcceptanceCriteria, TConstraints, TNonGoals with JSON serialization
  • Context models: TActor, TCollection, TConcept, TDocumentation, TReference, TStakeholder, TSubject with JSON serialization
  • Instruction models: TConvention, TSkill with JSON serialization
  • Memory models: TDecision, TEvent, TInsight, TMeeting, TProgress with JSON serialization
  • Meta models: TMetaData, TPromptable with JSON serialization
  • Root models: TActivity, TAgent, TChecklist, TContext, TGoal, TInput, TInstruction, TIssue, TMemory, TOutput, TPersona, TPromptField, TRole, TSpec, TTask, TTemplate, TTool, TWorkflow with JSON serialization
  • Spec models: TAbility, TFeature, TJourney, TRequirement, TScenario with JSON serialization
  • Tool models: TApi, TCli, TScript with JSON serialization
  • Workflow model: TStep with JSON serialization
  • Enums: TBodyType, TRefType
  • Core extensions: TCollectionExtensions for YAML and XML serialization of lists/maps
  • Core models: TConfig, TEmbedType, TMdSection, TRenderType
  • Spawn models: TFile, TFolder, TSpawnConfig
  • json_annotation dependency for generated JSON serialization

Changed #

  • BREAKING: Restructured from flat DTO-based architecture to domain-grouped workspace model system
  • BREAKING: Removed all previous DTO classes (TeamDto, AreaDto, RoleDto, CollectionDto, InstructionDto, etc.)
  • BREAKING: Removed example files and export configuration system (ExportConfig, ExportType, ExportResult)
  • Replaced path dependency on turbo_serializable with versioned constraint ^0.3.0

0.0.2 - 2026-03-19 #

Fixed #

  • Included generated .g.dart files in package

Changed #

  • Restructured into 6 domain groups with shared infrastructure

0.0.1 - 2026-01-01 #

Added #

  • Initial release
  • TurboPromptable base class extending TurboSerializable
  • Convenience name and description constructor parameters on all DTOs
  • Automatic FrontmatterDto creation from name/description
  • ExportConfig with fileType, bodyType, shouldExport, fileName
  • ExportType enum (md, yaml, json, xml, txt, dart)
  • BodyType enum (markdown, yaml, json, xml, dart)
  • Tree structure with config inheritance via resolveConfig
  • YAML frontmatter generation via generateFrontmatter
  • Body export via exportBody
  • FrontmatterDto for structured metadata (name, description, values)
  • ExportResult for export output (body, frontmatter, config, promptable)
  • export() method with shouldExport filtering
  • exportTree() method for recursive tree export
  • Hierarchy DTOs: TeamDto, AreaDto, RoleDto
  • Knowledge DTOs: CollectionDto, InstructionDto, WorkflowDto, ReferenceDto, TemplateDto, RawBoxDto, ActivityDto, AgentDto, RepoDto
  • Tool DTOs: ToolDto, ApiDto, ScriptDto, ToolMethodDto, ToolParameterDto
  • PersonaDto for agent identity
2
likes
150
points
325
downloads

Documentation

API reference

Publisher

unverified uploader

Weekly Downloads

Object-Oriented Prompting framework for defining AI agent prompts, roles, workflows, and tools as type-safe Dart objects.

Homepage
Repository (GitHub)
View/report issues

Topics

#prompts #agents #serialization

License

MIT (license)

Dependencies

json_annotation, meta, turbo_response, turbo_serializable

More

Packages that depend on turbo_promptable