Bundle
RTK command rewriting suggestions + tool output compaction for DSH. Suggests rtk-rewritten bash commands (deny-mode under rewrite) and compacts noisy tool output (bash/grep/read) to reduce context usage. Port of pi-rtk-optimizer for the Pi coding agent.
Bundle
MemoryHub (mh) integration for DeepSeek Harness: loads .memoryhub checkpoint memory on session start, saves dsh sessions back into checkpoints, and registers the mh workflow skill and tools
Bundle
Context-aware voice input for DeepSeek Harness with Web Speech, local SenseVoice transcription, model polish, editable Composer drafts, and user-controlled sending
Bundle
DeepSeek Harness dynamic context pruning plugin (model-driven context management)
Bundle
Zero-dependency MCP stdio server template with an out-of-the-box A-share (Chinese stock) data toolkit. Key-free, NDJSON protocol, drop-in dsh bundle.
Bundle
Focus board for DeepSeek Harness agents: durable, model-maintained notes in the session workspace that pin the objective, constraints, and decisions across compaction and sessions — with automatic context injection, an archive on clear, and an optional read-only web panel
Bundle
A search tool for DeepSeek Harness that learns to pick between quick and thorough search strategies using a contextual bandit (Thompson sampling).
Bundle
DSH web plugin: right-side context-and-files panel (directory browser, injected-context documents, read-only git commit graph) plus the session-log download action
Bundle
Windows MCP Hub for DeepSeek Harness: scan MCP servers configured in Claude Code / Codex / CodeBuddy, probe real connectivity via stdio & streamable-http initialize handshakes, enable/disable from a /mcp popup panel or the settings page.
Bundle
Thin on-demand context-retrieval layer for DeepSeek Harness: context_query / context_slice / context_grep tools over the sessionQuery and subprocess seams.
Bundle
DSH plugin for semantic code search via Milvus vector database
Bundle
Real-time weather, location, battery, and system-device context for DeepSeek Harness
Bundle
DeepSeek Harness 插件:把 DSH 的动态 Cordis 工具集(inspect/define/run/stop/undefine)以 MCP 暴露给 Claude Code。端点强制身份认证,token 可在 DSH 设置页配置。
Bundle
LongCat (LongCat-2.0) adapter for the DeepSeek Harness LLM seam — 1M context, thinking mode, tool calling
Bundle
会话级自动上下文压缩插件:回合中每步前 + 回合结束检测上下文用量,超过会话阈值自动 compact,摘要注入上下文后自然继续(DeepSeek Harness / DSH)
Bundle
DeepSeek Harness bundle for AI NetCafé's hosted tools. One install for the full catalogue, or pick one of four focused packs (tables, dev kit, doc flow, China facts) so you only pay the context cost of what you use.
Bundle
DSH 记忆管理插件(宿主级,标准化重写版):文件夹记忆库、印象、组合与双向链接、会话级注入计划与消息注入、规约记忆与最近会话总结的新会话自动注入、会话总结、跨工作区会话列表、6 个 Agent 记忆工具、左侧图谱双视图浮层与消息跳转、设置页开关。只依赖 DSH 核心扩展点(settings/tools/agent 事件/webServer),功能实现自包含,官方升级影响面最小。
Bundle
Out-of-tree dsh plugin: GitLab, GitHub, Gitee, Gitea, and Bitbucket tokens stay out of the model context, stored encrypted (AES-256-GCM) in a plugin-owned file; the model calls forge API tools on demand, and the web settings page manages sites and tokens.
Bundle
Conversation-local credential handoff that writes through the DSH credential service without returning the secret to model context.
Bundle
DeepSeek Harness plugin: cross-session memory for the model. Full-text search all past sessions (SQLite FTS5 via ctx.sessionQuery) and bring the strongest matching excerpts back into the current context — recall earlier decisions, commands, and mistakes.
Bundle
Cross-session memory cue bank for DeepSeek Harness: builds a persistent event-cue library (keywords, user language idiosyncrasies) and re-activates remembered event details on topic switches, mirroring human episodic recall.
Bundle
MCP 服务器可视化配置插件:在 DSH Web 设置界面管理 MCP 服务器(stdio / streamable-http / SSE),保存即生效,无需手改配置文件。| Visual MCP server management for DeepSeek Harness: configure stdio / HTTP / SSE servers from the settings UI, hot-synced without editing config files.
Bundle
Model-authored context pruning for DeepSeek Harness through the official compaction API.
Bundle
Curate tool output before it enters the model's context: keep error/warning lines and head/tail, spill the full text to disk with a pointer. Faster turns, smaller context, sharper attention — for every tool, not just bash.