dsh-task-effort

DeepSeek Harness Development & Runtime Low risk

Auto-adjust model reasoning effort per task (off/low/high/max): task classification, \[effort=...] markers, in-turn error escalation, DeepSeek peak-pricing-period capping with countdown notices, and manual-selection priority.

根据任务自动调整模型推理档位（off/low/high/max）：任务分类判定、\[effort=...] 显式标记、轮内出错升级、DeepSeek 高峰计价时段自动限档省钱并显示切换倒计时、GUI 手动选择优先。

How to install

DeepSeek Harness dsh plugin add dsh-task-effort

About

dsh-task-effort — 根据任务自动调整模型 effort DSH 插件：每次 LLM 调用前，根据**当前任务**自动决定模型的推理强度 reasoningEffort（DeepSeek 支持 off / high / max），不再全局固定一个档位。工作原理 DSH 的 agent 循环在每次模型调用前会走 agent/request waterfall（dsh-agent-loop 的 buildRequest），插件在瀑布的最外层改写 reasoningEffort，因此**最终生效**，优先级高于模型选择器/设置里的默认 effort。判定时机：agent/pre-step（每步执行前）读取本轮最新用户消息并分类，判定结果缓存到当轮结束；agent/request 时应用到请求配置。中途转向（steer / 追加消息）会触发重新判定。判定规则（优先级从高到低） 1. **显式标记**：消息里带 [effort=off|high|max]（也支持【effort:…】、 {effort:…}、effort=…；low→off，medium→high）→ 直接采用 2. **复杂任务关键词**（maxKeywords：重构/架构/分析/refactor/design/debug…）→ max 3. **继续类指令**（continuationKeywo…

GitHub ↗

Recommendation signals

46 Tool quality · Based on stars, downloads, maintenance, security and docs

– User interest · Adjusted by in-site views, install copies and download clicks

46 Overall

0views

0unique visitors

0install copies

0download clicks

0outbound clicks

Links

GitHub ↗ npm ↗ Report issue ↗

Basic safety check

Findings: None
Sources: curated:awesome-dsh-plugin.com, curated:awesome-dsh-plugin/awesome-dsh-plugin
Topics: deepseek-harness, dsh-plugin

Related plugins

MCP Server

Score100

langfuse

langfuse/langfuse

🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.

☆ 35.0K ↓ 6.7M Development & Runtime ↗

Claude Code

Score92

claude-code-templates

davila7/claude-code-templates

CLI tool for configuring and monitoring Claude Code

☆ 31.7K ↓ 10.7K Development & Runtime ↗

MCP Server

Score86

semiotic

nteract/semiotic

React data visualization library for streaming, networks, and AI-assisted development

☆ 2.7K ↓ 25.0K Development & Runtime ↗

DeepSeek Harness Featured

Score84

dsh-desktop

dataelement/dsh-desktop

DSHDesktop：DeepSeek Harness Desktop / DeepSeek Harness 桌面版

☆ 9.1K ↓ 930 Development & Runtime ↗