What problem does it solve? Setting up a consistent agent development harness across projects is tedious and error-prone, especially when a repository already has partial process directories, stale context files, or legacy layouts that must be preserved rather than overwritten. ## Core Features & Use Cases - Interactive Detection and Classification: Detects the project manifest (package.json, pyproject.toml, go.mod, Gemfile, Cargo.toml), package manager, framework, test runners, and monorepo structure, then classifies the repo as a new or existing project. - Approval-Gated Scaffolding: Creates the process/ directory tree from seed templates, migrates legacy layouts (plans/, reports/, skills/ dirs) with safe-inference rules, and never moves or overwrites files without user confirmation. - Deep Codebase Study: Runs parallel subagents to scan architecture, tests, context groups, and feature areas, then populates all-context.md and all-tests.md with real content instead of placeholders. - Use Case: You clone a repository that has a half-configured process/ directory from an old harness version. Run this Skill to study what exists, approve a migration plan, and end with fully populated, validated context files. ## Quick Start Ask the agent to run vc-setup to detect this project's stack and set up the agent harness, answering its questions as it goes.
