@jasonqq/dsh-btw-plugin
Codex-style /btw for DeepSeek Harness: ask a side question answered by a conversation-seeded subagent, keeping the answer out of the main context
99 results
Codex-style /btw for DeepSeek Harness: ask a side question answered by a conversation-seeded subagent, keeping the answer out of the main context
Give DeepSeek Harness agents the ability to read images directly: a model-facing read_image tool that answers questions about an image through any OpenAI-compatible vision endpoint.
System notifications for the DeepSeek Harness web GUI — OS-level popups for finished turns and pending interactions (approvals, questions, plan reviews). · DSH Web 界面系统通知:回合结束、待处理授权/提问/计划审核时弹出系统级提醒。
Codex Pets–style desktop pets + live task panel for DeepSeek Harness Web UI: a draggable anime pixel pet plus a floating dashboard of running/completed tasks and questions awaiting your input. / DSH Web UI 桌宠 + 实时任务面板(进行中 / 已完成 / 等待你回答的提问)。
Approval Assistant for DeepSeek Harness: approval reminders, auto-approval rules, completion reminders and question popups with per-feature channels (popup / system notification / sound / custom sound). 审批提醒、自动审批、完成提醒与提问提醒。
A floating right-side table of contents (TOC) for the DeepSeek Harness Web UI: one entry per user question, click to scroll to that message.
DSH 选择增强插件(生态唯一):给 ask_user_question / ask_user_choice 的选择卡加「重新生成选项 / 更多选项」两个按钮——选项都不满意一键换一批(换角度/维度,5~8 个),选项太少一键补充到 6~10 个。纯前端交互实现,不改核心包、不新增工具,原生文字卡与图片选择卡通吃。
Browser audio alert for human-attention events: approval requests, ask-user questions, and finished turns, with per-trigger sounds configurable through the audio-alert settings namespace
DeepSeek Harness dynamic plugin: renders three clickable, editable follow-up question suggestions under every agent answer
Semantic notification sounds for DeepSeek Harness: complete/approval/question/block/error via BEL or WAV, with Web UI bell toggle | DeepSeek Harness 语义化通知声音:完成/审批/提问/受阻/错误,经 BEL 或 WAV 播放,带 Web 铃铛开关
DSH Web 侧边栏锚定气泡:授权 / 提问 / 计划确认 / 目标阻断,会话行旁一键处理,无需点进会话
DSH web plugin: Codex-style /side side conversations. Open a temporary fork of the current chat in a floating panel — ask side questions without interrupting the main task or polluting its context.
Enforced pain-point-check guard plugin for DeepSeek Harness: after two non-converged experiments on one problem it injects the three questions, denies non-investigative tool calls until they are answered in the reply text, and blocks same-direction retries
SQL 优化模式:EXPLAIN 分析、索引策略、N+1 解决、查询改写。受 wshobson/agents(38k★ MIT)启发。
Universal local vision bridge tool for DSH: describe/answer questions about images via any OpenAI-compatible local inference framework (llama.cpp / Ollama / LM Studio / vLLM), vision models auto-detected
DSH web profile plugin: a right-side jump panel (like the DeepSeek web chat) listing the user questions of the current session, with one-click jump to the message and live reading-position highlight.
Turn DeepSeek Harness conversations into user-curated quiz questions.
Long-term memory for stock DeepSeek Harness: key/value store with question-time recall, plus optional migration from official ChatGPT, Claude, Gemini, and DeepSeek exports.
Question index surface: an ordered, collapsible list of the user's asked questions floating on the right side of the conversation, with click-to-jump into the flow
Semantic long-term memory for DeepSeek Harness: embedding-based retrieval over a persistent cross-session memory store, with model-facing tools, proactive per-question recall, automatic conversation summarization, and auto-selected embedding provider (apiKey present → API, otherwise local).
Fail-closed network recovery, recommended answers, and safe unattended continuation for DeepSeek Harness.
DSH <-> Telegram bridge as a profile plugin: forward session activity, replies, and question/approval button flows to Telegram, with a persistent GUI card in 插件配置.
DSH Web 插件:在会话中列出我的问题并点击定位
A DSH (DeepSeek Harness) bundle plugin that rewrites the ask_user_question tool description at prompt-assembly time with a fixed narrative-voice rule (default: the answerer narrates), toggled live via the /voice command.