dsh-refusal
Run the following command in DeepSeek Harness:
dsh plugin install uckkk/dsh-refusal
Paste the following prompt into your AI chat to install this plugin:
Install the plugin in DeepSeek Harness by running dsh plugin install uckkk/dsh-refusal; the full source code is available at https://github.com/uckkk/dsh-refusal .
About this plugin
When a language model has to turn down a request, a flat "I cannot do that" rarely feels warm, and it almost never tells the user what to try next. dsh-refusal turns refusal into a structured output: it returns context-appropriate rejection methods together with actionable alternative suggestions, so saying no also carries tone, clarity, and a concrete path forward.\n\nThe plugin registers a single tool called refusal. Once invoked, it reads the scenario you pass in and produces a structured set of rejection approaches and alternative suggestions. Everything is written in pure Node with zero network calls, which means it runs comfortably in offline environments, intranet gateways, or resource-constrained containers without any external dependencies to keep up to date.\n\nIf you are building a customer-service bot, a content-safety gateway, or any product that needs to follow a refusal with a concrete next step, dsh-refusal offers a lightweight, plug-and-play option. Once installed, you can call the tool directly inside a session with no extra configuration, no API keys, and no network dependency to worry about.
Use Cases
- Customer-service bots that need a polite refusal plus alternatives
- Content-safety gateways returning structured rejections after a block
- Offline or intranet deployments requiring a zero-network refusal module
Best For
- Developers building conversational AI products
- Engineers on content-safety and moderation teams
- Engineers deploying AI services in offline or restricted networks
Related Plugins
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.