Agent Name

MVP Scope Guard

Short Description

Turn rough software ideas into bounded MVP scopes, practical architecture decisions, and coding-agent-ready implementation tasks without unnecessary complexity.

Expertise Behind This Skill

This Skill is built from practical software product planning and AI-assisted development workflows. It focuses on preventing feature creep, oversized tasks, premature infrastructure, and documentation that does not match the actual implementation plan.

Capabilities

Classifies ideas as Prototype, Small MVP, Product MVP, or Platform
Identifies the primary user and primary user outcome
Separates confirmed requirements from assumptions
Defines must-have, deferred, and explicitly excluded scope
Selects the simplest suitable architecture
Detects premature microservices, Kubernetes, multi-tenancy, event-driven architecture, and other unnecessary complexity
Generates project documentation for human developers and coding agents
Breaks implementation into focused tasks with acceptance criteria
Performs a final scope consistency audit

Output

The Agent produces a Markdown-based planning package containing:

PROJECT_BRIEF.md
README.md
AGENTS.md
ROADMAP.md
DECISIONS.md
SCOPE_AUDIT.md
Sprint documentation
Coding-agent-ready task files

Use Cases

Turn a product idea into an executable MVP plan
Prepare a new repository for an AI coding agent
Reduce an oversized software proposal
Convert informal requirements into development tasks
Check whether a proposed architecture is unnecessarily complex
Create consistent project planning documents

Example Input

I want to build a personal expense tracker for Windows. I am the only user. It needs expense entry, categories, monthly totals, and CSV export. I want a usable first version within two weekends.

Example Result

The Skill classifies the project as a Small MVP, removes unnecessary authentication and cloud infrastructure, recommends a simple local architecture, and produces a bounded implementation plan with project documentation and focused development tasks.

Limitations

Does not implement the complete application
Does not guarantee delivery timelines
Does not replace security, legal, accessibility, or production architecture reviews
Does not automatically deploy infrastructure
Results depend on the accuracy and completeness of the user's requirements

FAQ

Does this Agent write the application?

No. It creates the bounded project plan, documentation, architecture decisions, and implementation tasks that a developer or coding agent can execute.

Does it support existing projects?

It is primarily designed for new projects and scope reviews. It can review an existing proposal, but it is not a code-review or bug-fixing Agent.

Will it always recommend the simplest technology?

It recommends the simplest architecture that satisfies confirmed requirements. More complex infrastructure is included only when a concrete requirement justifies it.

Can I use the result with coding agents?

Yes. The generated files are designed to be readable by human developers and common coding-agent clients.