#9595th of 98 in Skills, plugins and rules
dsh-routing-suite
Plugins for the DSH agent: runtime injector, routing presets, graded task protocol
- Stars
- 7k
- License
- MIT
- Last commit
- Sep 2026
- Last release
- Sep 2026
- Language
- JavaScript
Missed the latest GitHub check; these facts may be older.
Overview
Bundles three add-ons for the DSH agent (@deepseek-ai/dsh): an injector that adds dev_* tools for hot-reload without restart, router-standard and router-spec presets that pick a reasoning persona per task type, and a graded two-level task protocol with a red-team gate and audit endpoint. Installed with dsh plugin add or install.ps1. Targets DSH >=0.1.0-rc.6 <0.2.0, a developer preview with announced breaking changes.
Who it is for: DSH users wanting runtime tool injection, task routing and a task protocol
Strengths
- Three components in one repo, installable with a single dsh plugin add command
- Includes measure.mjs to recompute the reported metrics on your own session logs
- Experiment reports (P1-P23) and measured data are published alongside the presets
- Presets replace only the persona section, so plan-mode boundaries are kept
Weaknesses
- Works only with DSH, a developer preview with announced breaking changes
- Install steps and scripts are PowerShell; other platforms are not documented
- Main README and docs are in Chinese, with a separate English README
- router-pro preset is planned and not released
What it needs
- Needs DSH (@deepseek-ai/dsh) >=0.1.0-rc.6 <0.2.0, Node.js, PowerShell
- Models: DeepSeek Pro, DeepSeek Flash
Also in Skills, plugins and rules
See all 98| Rank | Project | Score |
|---|---|---|
| 1 | skillsAgent skills for planning, TDD, debugging and code review | 94 out of 100 |
| 2 | eccAgents, skills, hooks and rules bundle for Claude Code, Codex and other harnesses | 94 out of 100 |
| 3 | cavemanCuts token use in coding agents with terse replies and compressed inputs | 94 out of 100 |
| #4 | impeccableDesign skill with 24 commands and a detector for AI-generated frontends | 93 out of 100 |
| #5 | ponytailSkill prompt that makes coding agents write less code and add tests | 92 out of 100 |