dsh-suggest
Run the following command in DeepSeek Harness:
dsh plugin install shaoeric/dsh-suggest
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install shaoeric/dsh-suggest in the DeepSeek Harness terminal to install the plugin from https://github.com/shaoeric/dsh-suggest.
About this plugin
When iterating with an agent across multiple turns, the most common friction is knowing what to ask next after an answer arrives. dsh-suggest closes that gap by rendering three contextual follow-up suggestions directly beneath every completed response, turning a passive pause into an active choice of direction.
Each suggestion is anchored to the three most recent user messages (including the assistant replies between them) and generated via the same provider's flash model with reasoning disabled, so results arrive quickly at low token cost. Users can click a suggestion to send it immediately, or hover to reveal an edit button that drops the text into the input box for refinement before dispatch. Failures surface a specific reason with a one-click retry, and results are cached per session and turn to eliminate redundant calls.
Suggestions appear only under the latest turn and disappear once a new exchange begins, preventing accidental clicks on stale prompts from earlier in the conversation. All colors follow DSW theme variables for seamless light and dark mode support. The plugin suits DSH users who want faster multi-turn iteration, lower typing overhead, and a “scan then commit” workflow when exploring with DeepSeek agents.
Use Cases
- Answer is in but you are unsure what to ask next
- Rapidly exploring multiple angles of the same topic
- Cutting down repetitive typing of follow-up questions
Best For
- Developers who iterate heavily with DSH agents
- Researchers seeking lower input cost for follow-ups
- Explorers who prefer a scan-then-commit workflow
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.