qjf44/dsh-plugin-thinking-api ↗★ 0
dsh-plugin-thinking-api
DeepSeek Harness 插件：一键配置任意 OpenAI 兼容 API 并自动带思考模式，同时规避 developer 角色被第三方端点拒绝（content_filter）的问题。
AI 分析
核心用途是简化第三方大模型（如DeepSeek-R1等带思考模型）的接入。适合需要使用第三方API、且希望完美呈现思考过程的开发者。
インストール
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:qjf44/dsh-plugin-thinking-apiドキュメント
README 全文を読む ↗Configuration reference
Provider (per providers.)
| Field | Type | Required | Default | Meaning |
|---|---|---|---|---|
baseURL | string | ✅ | — | API endpoint base URL |
apiKeyEnv | string | — | — | Env-var name holding the API key |
displayName | string | — | route id | Name shown in pickers |
thinkingFormat | enum | — | deepseek | deepseek | openai | openrouter | together | zai | qwen | string-thinking |
models | dict | ✅ | — | Model id → model entry |
Model (per models.)
| Field | Type | Required | Default | Meaning |
|---|---|---|---|---|
name | string | — | model id | Display name |
thinking | boolean | — | false | Enable reasoning levels |
thinkingEfforts | dict | — | auto | Custom level → wire-value mapping, e.g. { off: null, high: high, max: xhigh } |
contextWindow | number | — | 262144 | Context window size |
maxTokens | number | — | 32768 | Max output tokens |
When thinking: true and no thinkingEfforts is given, the plugin fills a verified DeepSeek-compatible level map (off / high / max, mapping to disabled / reasoning_effort: high / reasoning_effort: xhigh). Provide thinkingEfforts to add levels (low/medium) or override per API.