What problem does it solve?
This Skill removes the friction of starting AI-assisted work on an existing or new codebase by automatically discovering project structure, detecting the tech stack, and establishing a clean shared project index. It helps teams avoid repeatedly explaining repository layout, languages, frameworks, and project boundaries every time work begins.
Core Features & Use Cases
- Project discovery and structure scanning: Finds root-level package files, source folders, test directories, config files, and sub-projects to determine whether the repository is a single project or monorepo.
- Tech stack and project type inference: Detects primary language, secondary languages, frameworks, build tools, and test frameworks, then classifies each project as frontend, API service, CLI tool, library, monorepo, or generic.
- Safe initialization and refresh workflow: Supports first-time setup and refresh mode, asks for confirmation before overwriting existing artifacts, validates generated YAML, and preserves selected existing context during re-scan.
- Use case: When onboarding MVTT into an unfamiliar repository, use this Skill to build the initial .ai-agents workspace index so later analysis, design, and implementation skills can operate with reliable project context.
Quick Start
Ask the AI to run mvt-init on the repository to scan the codebase, confirm detected projects and tech stack, and write the initial MVTT project context.