Per-message delete for DeepSeek Harness: remove a user message, one reply step, or a whole assistant reply from the derived model context via the official surface-replace contract, and hide it from the visible transcript without rewriting the append-only log.
Windows desktop UI context picker for AI agents: select windows, UI elements, or screen regions with hover highlighting, UI Automation, screenshots, local OCR, and user-approved MCP output. Works with OpenAI Codex, DeepSeek Harness, and other MCP-compatible clients.
dsh侧边HUD面板,包含多个信息展示模块(可自定义是否展示),集成压缩上下文、查看git graph等功能。DSH side HUD panel, containing multiple information display modules (customizable whether to display), integrating features such as compressed context, viewing git graph, etc.
DeepSeek Harness native plugin: Settings UI for the Roblox Studio MCP bridge - list connected Studio instances, pick the active one, and expose the selection to agents as runtime context.
Open a persistent side conversation from any DeepSeek Harness session to investigate questions, clarify requirements, or explore alternatives without polluting the main task’s context, then review and send a concise conclusion back to the parent conversation. Tested with DSH 0.1.2-alpha.1.
A full engineering bench for your coding agent: the work survives any session, and every done comes with proof. Runs on Claude Code, Codex CLI, Copilot CLI, Kiro CLI, and DeepSeek Harness.
Scopes MCP tool injection per workspace directory in the DeepSeek Harness: a session opened in a project sees only the MCP servers that project needs — removed from the model's tool list and refused at call time. Plus per-session switches in the composer to narrow or widen the session you are in, temporarily.
Zero-dependency MCP stdio server template + ready-to-use A-share data tools for DeepSeek Harness (dsh). No API key, all env-var paths, standard NDJSON protocol.
Context window optimization for DeepSeek Harness agents. Sandboxes tool output in-process (only what you print enters the conversation), persists session memory in an FTS5 knowledge base, and enforces routing via a system-prompt routing section.
DeepSeek Harness plugin: GitLab and GitHub API tokens stay out of the model context — encrypted at rest (AES-256-GCM), tools on demand, web settings panel.