Repository fingerprint
Detect repo shape, language, frameworks, package managers, manifests, commands, environment signals and important code areas.
Repo2Skill maps architecture, commands, dependencies, APIs, tests, CI and code structure, then packages two repository-native Skills: ChatGPT and Claude.
Repo2Skill reads the repository tree and a prioritized set of high-signal files directly from GitHub in your browser.
Public repositories only in this zero-cost build.
Review what Repo2Skill actually detected before compiling the Skills.
Repo2Skill reads the repository tree and a prioritized set of high-signal files directly from GitHub in your browser.
Detect repo shape, language, frameworks, package managers, manifests, commands, environment signals and important code areas.
Extract routes, exported symbols, tests, CI, dependencies, conventions and source excerpts with exact repository paths.
Compile implement, debug, review, test, explain, refactor and migration protocols with validation and source-of-truth rules.
No. This build uses deterministic browser-side analysis of GitHub metadata, the recursive tree and selected raw files. That keeps recurring AI API cost at zero.
It builds a bounded evidence snapshot, not a perfect semantic index. Deep mode fetches more high-signal files and includes a file index plus source excerpts. When a live repository is available, the Skill is instructed to prefer current files over the snapshot.
The Skill is generated from the repository’s real commands, architecture signals, dependencies, routes, symbols, tests, CI and conventions. It also has task-specific protocols for implementation, debugging, review, testing, refactoring and migrations.