What problem does it solve? Replacing an application without changing its specification requires evidence-based decisions, not guesswork. This Skill acts as the entry point for a behavior-preserving replacement: it measures the current application (semantics, DB restore capability, side effects, file I/O reachability), derives a replacement strategy from measured values, decomposes the app into features, and dispatches work to sister skills (golden-dataset / parity-suite / parity-replace / parity-diff). It never implements code itself and stops when measurement is impossible. ## Core Features & Use Cases - setup mode: Checks dependencies (issue-create, browser-test, chrome-devtools MCP), runs an interactive project setup, measures the current app with a role probe script, records a strategy, builds a feature inventory, and decides shared package dependencies. - issues mode: Selects unfiled features from the inventory and files GitHub Issues via the issue-create skill, with an explicit approval gate before any external side effect. - status mode: Derives current progress and unverified areas from GitHub Issues and in-repository artifacts without holding its own state. - Use Case: You want to migrate a legacy order-management app to a new TypeScript stack. Run setup to measure the current app, get a parity-test strategy and feature inventory, then file per-feature Issues that downstream skills implement and verify with deterministic diffing. ## Quick Start Ask the AI to run replace-strategy setup to measure the current application and produce a replacement strategy and feature inventory.