Repository navigation
VibeLens v0.15.0
Model independent workflows and object lifecycles in one project, with a smaller
Skill download and verified viewer setup.
Changes
- Register multiple state models and named subnets in each layer. Preserve
independent object identities, shared physical declarations and explicit
cross-model relationships across 2D, 3D, live observation and Timeline. - Carry model identity through native Python/JavaScript execution, snapshot v2,
collection, replay and mapping verification. Initial discovery and iteration
recall include independent business capabilities and state-owning objects. - Move 37 duplicated third-party files out of tracked source, removing 23.6 MB.
Fetch pinned resources from a CDN first and GitHub second; verify every file
by SHA-256 and size. Reuse verified cache for later/offline launches. - Prepare resources during source installation or first viewer use. Preserve
modified dependencies, reject incomplete managed upgrades, and provide an
optional offline package with corresponding source and notices.
The portable native executors still enforce their strict flat-state subset.
Multiple models do not imply support for arbitrary hierarchy, concurrency or
cross-model transactions. Original library licenses remain unchanged.
Install
npx skills add li-bruce/VibeLens --skill vibelensUse $vibelens in Codex or /vibelens in DeepSeek Harness. Python 3.11+ is
required. First viewer use downloads about 6.75 MB of pinned resources; the
Skill guides this step. Existing projects keep their native state/probe adapters
and use the reviewed upgrade workflow when adopting new model capabilities.
vibelens-v0.15.0.zip: Skill code, references, notices and dependency lock.dsh-vibelens-0.15.0.tgz: the same Skill as a Harness plugin.vibelens-skillhub-v0.15.0.zip: the catalog transport package.vibelens-offline-v0.15.0.zip: Skill plus all pinned viewer resources.SHA256SUMSandmanifest.json: archive hashes and owned-source inventory.