v0.1.0 - Umbra, the apophatic code auditor

First tagged release of Umbra, the apophatic code auditor.

Umbra reports what your tests are structurally unable to see, and proves each gap with a real failing test. It maps what a test suite actually exercises, finds the reachable and meaningful cases nothing asserts, and reports a Silence index: the honest complement to coverage. Higher means more blind spots. Then it writes failing tests that close the gap. For Python, TypeScript, JavaScript, Java, Rust, and Go.

It runs two ways: as a Codex plugin that audits code the moment the agent writes it, and as a standalone CLI.

• Demo (2:58): https://www.youtube.com/watch?v=WQa6vukBqKQ
• OpenAI Build Week submission: https://devpost.com/software/umbra-reveals-what-your-tests-aren-t-telling-you
• Umbra is the module-level instrument in the Slop Audit, stewarded by the Open Honest Foundation: https://github.com/openhonest/slop-audit

Built with Codex and GPT-5.6.