What problem does it solve? Before starting a long novel-generation run, a misconfigured provider (stopped local proxy, expired API key, wrong base_url) only surfaces when the first real call fails. This Skill runs a minimal real Generate round-trip against the default model and every role model (coordinator / architect / writer / editor), reporting per-target availability so configuration problems are caught before writing begins. ## Core Features & Use Cases - Per-target connectivity check: Deduplicates by provider/model and pings each target once, reporting latency or the exact error for both primary and configured fallback models. - Per-role availability summary: Reports whether each named role is usable, degraded to fallback, or fully down, with a non-zero exit code when any role has no working path. - Ad-hoc provider validation: Test a spare provider/model pair without changing configuration to confirm it can serve as a fallback. - Use Case: Before launching a multi-chapter generation pipeline, run the check to discover that the local proxy for one provider is not running, fix the base_url, and confirm all roles pass. ## Quick Start Ask the assistant to run the novel-studio connectivity self-check to verify that the default model and all role models are reachable before starting creation.
