What problem does it solve? Verifying that kajet-turbo actually works at runtime requires a database, an embedding provider, and a running server, which is tedious to wire up by hand. This Skill provides a repeatable launch recipe and a sequence of REST calls that exercise login, workspaces, notes, jobs, and reindexing against a fully disposable environment. ## Core Features & Use Cases - Isolated launch recipe: Boots the app with a temp SQLite database, temp workspace directory, and a fake OpenAI-compatible embeddings endpoint on port 8899, with no production dependencies. - REST surface walkthrough: Drives login, workspace creation, embedding profile setup, and note creation via curl, including the expected_sha requirement for PATCH/DELETE operations. - Runtime observability: Explains how to inspect JSONL logs, query the jobs API, and count vectors using sqlite_vec since the sqlite3 CLI cannot read vec0 virtual tables. - Use Case: After modifying the note chunking or embedding pipeline, run this Skill to confirm notes are created, embedded, and indexed correctly before committing. ## Quick Start Ask the AI to verify kajet-turbo at runtime by launching it with a temp database and fake embedding endpoint, then driving the REST API to create a workspace and a note.
