Practical release security.
Clear about the limits.

What CloneGuard does today

The v0.1 skill and Python CLI audit local artifacts, stage loose web releases with known private files excluded, create release receipts and guide platform-native hardening. They do not make client software unreadable, and have not established that REA cannot reconstruct a protected application.

The public repository contains the implementation, tests and recorded synthetic results. The website source is private; the deployed guides and product information are public.

How this library is produced

These articles were prepared with AI assistance to turn official toolchain documentation and recorded project checks into practical release workflows. Technical references are linked on each page. They are not presented as independently reviewed research or evidence that every procedure has been tested on customer software.

We separate general guidance, planning and measured observations. A research note identifies its pinned tools and recorded evidence where available. A 2027 guide is a forward-looking plan written in October 2026. Publication dates are not advanced to imply freshness, and build time does not change article dates.

We do not publish invented attacker hours, token usage, cost multipliers or guaranteed protection. A hash receipt measures artifact changes; an analyzer budget limit measures a bounded probe. Neither establishes an inability to clone a product.

If a technical detail is wrong or out of date, open a correction issue with a public reference or a synthetic example. Do not submit private code, secrets or unpublished customer evidence.

Application reviews

A scoped review can examine release artifacts, build settings, public source maps, diagnostics and sensitive API boundaries. Deliverables depend on the agreed scope and available platforms. No claim is made that a public product can become impossible to copy.