Skip to main content
Version: 0.4.x

What "AI-Native" Means​

"AI-native" does not mean adding an AI assistant button to your product, nor does it mean using AI to help with one specific feature. AI-native means making AI the primary path of the product, not an optional add-on. In an AI-native design, AI is not a tool — AI is the team's core engineering productivity. The developer's role shifts from code writer to direction-setter and key decision-maker.

LinaPro was designed from the ground up with AI-driven development in mind:

  • It provides a spec-driven AI-native development workflow, allowing developers to hand off requirements analysis, system design, code implementation, test verification, and other development work to AI.
  • It also ships a built-in AI skill system covering the full development lifecycle, giving AI the domain-specific knowledge to make framework-compliant decisions in concrete work contexts such as backend development, frontend design, testing, and performance auditing.

These two dimensions complement each other and together form the foundation of LinaPro's AI-native design.

AI-Native vs. AI-Assisted​

DimensionAI-NativeAI-Assisted
AI's roleCore engineering productivity, primary pathOptional support tool
Where code comes fromAI writes, humans decideHumans write, AI assists
DocumentationAI maintains it in sync, tightly coupled to codeHumans maintain it, prone to drifting from code
Architecture consistencySpec-anchored, AI automatically followsFully dependent on manual review
Test coverageMandatory E2E tests, a prerequisite for every changeWritten as needed, gaps possible
Iteration speedBounded by AI processing speedBounded by human coding speed

AI-Native Development System Design​

AI Across the Entire Development Lifecycle​

In LinaPro's AI-native workflow, AI is deeply involved in every stage of development:

  • Requirements Analysis: AI reads project specifications (AGENTS.md), existing OpenSpec documents, and source code to deeply understand the current architecture, proactively raises clarifying questions, and helps you think through boundaries and impacts.

  • System Design: AI generates database table designs, API interface definitions, and frontend page structures, producing complete design documents (design.md) and incremental specs (specs/) for your review and decision-making.

  • Code Implementation: AI works through the task list (tasks.md) item by item — backend services, API routes, database DAO layer, frontend page components, permission declarations, and i18n resources — all in one go.

  • Test Verification: AI writes E2E test cases, runs tests, and ensures the implementation matches the design specification.

  • Documentation Archiving: AI organizes design decisions and interface specifications from the change into baseline specifications for future iterations to reference.

Built-in AI Skill System​

AI-native means more than just a development workflow — it means the framework provides deeply customized skill (Skill) support for every stage of AI's work. LinaPro ships dozens of built-in AI skills covering the full development lifecycle, enabling AI to work at peak efficiency in every specific scenario without having to repeatedly explain project conventions in each conversation. Here is the complete built-in skill system:

CategorySkillCore Capability
AI Workflowopenspec-*Spec-driven AI-native development workflow, covering exploration, proposal, implementation, and archiving
Project Governancelina-reviewCode and specification compliance review, quality gate before OpenSpec changes
Project Governancelina-feedbackUser feedback triage and execution loop, issue tracking, bug fixes, regression testing
Project Governancelina-e2ePlaywright E2E test case naming, organization, and authoring conventions
Project Governancelina-perf-auditFull backend API performance audit, detecting N+1 queries, missing indexes, and other performance risks
Archive Managementlina-openspec-archive-changesScans and archives all completed active OpenSpec changes
Archive Managementlina-openspec-archive-consolidateAggregates and summarizes archived changes by functional responsibility
Frontend Developmentfrontend-designProduction-grade frontend UI design, avoiding generic "one-size-fits-all" AI interface styles
Frontend Developmentfrontend-patternsFrontend development patterns, state management, performance optimization, form handling best practices
Frontend DevelopmentvbenVben Admin 5.0 frontend framework development guidance, covering routing, permissions, themes, etc.
Community Automationlina-community-*Automated handling of community tasks such as GitHub Issues and PR reviews
Browser Automationplaywright-cliBrowser automation for E2E test execution and debugging
Version Managementgit-commit-pushAnalyzes diffs and auto-generates repository-compliant commit messages, then pushes
Version Managementgit-worktreeCreates isolated Git worktrees for parallel tasks to avoid interference
Coding Standardskarpathy-guidelinesAvoids common LLM coding pitfalls, ensuring precise implementation without over-engineering

These skills are embedded in the framework's AI collaboration specifications (.agents/skills/) as domain knowledge. Built-in project skills load automatically with the source code, and AI tools automatically activate relevant skills when handling corresponding scenarios. External tools or skills such as OpenSpec CLI, goframe-v2, and find-skills are still recommended for installation on demand to get the complete spec-driven and backend development experience. The direct benefit of a rich skill ecosystem is: AI can make decisions that comply with the LinaPro framework's constraints in every specific work scenario, rather than relying on the model's general knowledge to generate code or documentation that doesn't conform to project conventions.

Progressive Project Specification Management​

LinaPro does not cram all rules into one massive prompt. Instead, it adopts a progressive project specification management strategy. The AI Agent first reads the project's top-level entry point, then continues reading detailed rule files based on which domains the current task touches.

Specification ResourceResponsibility
AGENTS.mdTop-level specification entry, declares project positioning, mandatory trigger scenarios, rule loading gates, and rule file index
.agents/rules/*.mdDomain-specific rules split by area: documentation, plugins, backend, frontend, database, testing, development tools, etc.
.agents/skills/Built-in project skills, enabling AI to use specialized workflows and constraints in specific scenarios
.agents/prompts/Project commands and prompt directory, e.g., OpenSpec-related command entries
apps/lina-plugins/<plugin-id>/AGENTS.mdPlugin-local specification entry, applies only to file changes within that plugin directory

Plugin-local specifications are a critical part of this strategy. When developing a plugin, if the plugin directory contains an AGENTS.md, the AI Agent must prioritize reading it. When plugin-local specifications conflict with the project's top-level specification or .agents/rules/*.md, the plugin-local specification takes precedence within that plugin directory. Uncovered portions continue to follow the global conventions and matched rule files.

The value of this design is that the main project maintains a unified engineering quality baseline, while plugins can express their own domain constraints, directory conventions, business rules, and testing requirements. As the number of plugins grows, AI does not need to load all plugin rules at once — instead, it reads the corresponding local specification when entering a specific plugin's development context.

To accommodate different developers' preferred AI Coding tools, LinaPro provides make agents series commands that bridge the unified source to each tool's own convention paths. For example, tools that natively read AGENTS.md need no extra files; tools that only read CLAUDE.md, GEMINI.md, or QWEN.md can use the same root specification via symlinks. Skills and prompts are similarly bridged to .claude/skills, .codex/prompts, .cursor/commands, and other paths.

Specification-Driven Development​

LinaPro's AI-native workflow is built on the Specification-Driven Development (SDD) philosophy.

The problem with traditional development is that over time, documentation falls behind code, code drifts from design, test coverage develops gaps, and eventually no one dares touch the "legacy code."

SDD solves this through the following mechanisms:

Every Change Has a Corresponding Spec Anchor​

Each feature iteration produces complete documentation in the openspec/changes/ directory, including design decisions, interface specifications, and implementation details. These documents serve as the context foundation for AI in the next iteration.

Specification First, Code Follows​

Before writing a single line of code, AI must first establish the specification (proposal.md and design.md pass review), and only then begin implementation. This ensures code and documentation are produced in sync within the same iteration cycle.

Mandatory E2E Tests as a Prerequisite for Change​

Every change must have corresponding E2E test coverage. Tests passing is a hard prerequisite for archiving a change, not an optional step. This fundamentally prevents the accumulation of "test gaps."

Archiving Is Crystallization​

When a change is archived, incremental specifications are merged into the openspec/specs/ directory, forming a continuously growing baseline specification library. AI uses these verified baselines as constraints in subsequent iterations, ensuring architectural consistency.

This is a question many developers have when they first encounter LinaPro: can I just modify code directly without going through the AI workflow?

You can, but it's not recommended. Here's why:

Documentation and Code Will Immediately Drift Apart​

LinaPro uses SDD specification documents (design.md, specs/) to describe the system's current state, managed through OpenSpec. If you modify code directly without updating the specifications, the next AI iteration will generate designs based on outdated specifications that don't match reality, leading to increasing divergence.

AI Cannot Perceive "Unspoken Rules"​

When humans modify code directly, they often introduce implicit constraints (a special condition added here, a limitation bypassed there) that exist in the developer's mind but not in the specification documents. When AI modifies the same module next time, it will likely break these implicit constraints.

AI Can Guarantee Implementation-Documentation Consistency​

When AI drives implementation, code and documentation are produced simultaneously in the same conversation, ensuring they are always consistent. This is an efficiency that human work patterns can rarely achieve.

Special Cases​

Direct code modification may be considered in these scenarios:

  • Urgent production issue fixes (but OpenSpec records must be supplemented afterward)
  • Highly specialized business constraints that AI cannot understand
  • Configuration file and environment variable adjustments

Even in these scenarios, it is recommended to use the /lina-feedback skill afterward to record the changes in the current active OpenSpec change.

The Value of an AI-Native Framework​

The core value of LinaPro's AI-native design can be summarized as:

  • Delivery Speed: AI can complete a full CRUD module in minutes, including backend API, database layer, frontend pages, and test cases — work that would take a human engineer at least several hours.

  • Quality Consistency: Every iteration follows the same conventions and best practices. Code style, naming conventions, error handling, and test coverage remain highly consistent, unaffected by individual developer habits.

  • Sustainable Evolution: Specification documents stay in sync with code updates. Architectural decisions are documented and traceable. New members can quickly understand the system's design intent by reading openspec/specs/.

  • Risk Control: Every change has E2E tests as a safety net. The AI review mechanism catches deviations from specifications in time, eliminating issues before they reach the main branch.

  • Scenario Awareness: Over a dozen built-in AI skills cover the full development lifecycle. Whether it's backend API development, frontend UI design, E2E test authoring, performance auditing, version upgrades, or code commits, AI has domain-specific knowledge for every work scenario. Developers don't need to re-explain project conventions to AI each time, significantly reducing the cognitive friction of AI collaboration.