ice-kele/dsh-llm-guardian
ice-kele★ 0JavaScriptLast synced: 2026-08-16
Provider health, quota, account usage, and local model usage statistics for DeepSeek Harness
README excerpt
DSH LLM Guardian 简体中文 English dsh-llm-guardian is a community plugin for DeepSeek Harness. It adds provider health checks, local token quotas, and provider-scoped account or plan usage queries to the Models settings page without storing API keys in the plugin settings. Provider cards The plugin contributes a health-test action, a model-usage action, and a compact status summary to every configured provider card. Local usage dashboard The optional dashboard aggregates local session logs into 7, 30, or 90-day views. It includes model filtering, summary metrics, an activity heatmap, a responsive daily token chart, model share, and detailed hover values. Features - Periodic and on-demand provider health checks. - Local token counters and optional request blocking after a configured quota is reached. - Provider-scoped account or plan usage queries with explicit timeout and refresh controls. - Built-in query adapters for DeepSeek balance and Z.AI Coding Plan quotas. - Local session-log statistics with model filtering, activity heatmap, daily token trends, and model share. - Secret redaction, response-size limits, and same-origin restrictions for custom usage queries. Install from GitHub …
View full README on GitHub →