dsh-side-chat
DeepSeek Harness 原生并排侧聊:归档独立会话、按需主会话上下文、原生会话 UI 与可配置生命周期。
98 results
DeepSeek Harness 原生并排侧聊:归档独立会话、按需主会话上下文、原生会话 UI 与可配置生命周期。
DeepSeek Harness plugin: serve LLM calls with the OAuth tokens your local Codex CLI and Claude Code already hold, instead of a separately configured API key — plus a composer badge showing subscription usage for those routes and for GLM Coding Plan.
DSH plugin: start/stop a local llama.cpp llama-server from the settings → plugins page — two model slots (A/B) with per-slot model-file selection, eight editable launch-parameter groups (text/vision × fast/long × 2 slots), and one shared DSH provider carrying both slots' models with four thinking-effort levels.
Enhanced subagent delegation tools for DeepSeek Harness with per-call cwd control: everything in dsh-subagent-tools plus a cwd parameter, shipped with the two in-process provider patches it requires.
Unified web search provider for DeepSeek Harness (ctx.web): 11 backends — Exa + Parallel + DDG + SearXNG (keyless), DeepSeek + Anthropic + OpenAI + Brave + Tavily + Serper + Mojeek (key-gated via Settings credentials). Registers provider id "unified" for zero-config web_search tool.
Installable DeepSeek Harness profile bundle for a separately installed MediaCrawler checkout
GrayPrint — live reasoning-style fingerprints for the DeepSeek Harness web chat. Compares grayscale-vs-current writing style by internal reasoning paragraphs and independently tracks Let me exploratory openers versus We need planning openers.
Register AllRouter as a DSH provider: one base URL and one key for Kimi K3, Claude, GPT, Grok, GLM, DeepSeek, MiniMax-H3. Switch models without managing separate providers; every call reports real cost.
DSH plugin — Hermes-style Mixture-of-Agents request aggregator: router → parallel reference advisors → aggregator stream, with dual-provider failover (official + OpenCode Go) and V4P triple-gate cost guard.
OpenCode-style web search (Exa / Parallel MCP) provider for the DeepSeek Harness ctx.web seam. Gives any model, including opencode-go's deepseek-v4-flash, a working web_search tool without a DeepSeek official API key.
Durable Git worktree isolation and delivery for DeepSeek Harness subagents
DSH web plugin: local offline FunASR voice input (paraformer int8 onnx sidecar, Web Speech fallback, LLM polish).
DSH plugin (NInfer engine only): fixes compaction failure on local qwen3.8-27b gateways served by NInfer — xhigh thinking burns the entire output token budget, so thinking is off for compaction-only, with the model's non-thinking sampling parameters; the same idea applies to other launch methods
Configurable Codex-style conversation navigator for the DeepSeek Harness Web GUI, with compact previews and fast paragraph jumps.
Task Flow / Execution Graph plugin for DeepSeek Harness (DSH): visualize a single task's full execution — turns, LLM steps, tools, skills, subagents, retries, parallel branches — as an interactive DAG with live updates and trajectory linkage.
Role-based AI swarm orchestration for DeepSeek Harness: duty-table model pinning per role, parallel task DAG with review loops, and a live Swarm dashboard tab in the web GUI.
3D printing & additive manufacturing capabilities for DeepSeek Harness: parametric STL generation, calibration G-code, STL/G-code analysis and toolpath rendering as model tools, plus slicing / modeling / analysis / diagnosis knowledge skills.
Complete Git worktree support for DeepSeek Harness: task-scoped isolated worktrees with per-repo manifests, workspace registration, bring-back to main, direct commit, and durable state. Reference: Qoder / Codex / Claude Code worktrees.
DSH web-ui plugin: '/' glyph on the composer command-menu button plus a '+' attach button with separated image/document/other uploads.
Merged Feishu/Lark plugin for DeepSeek Harness: Lark channel bridge (long-connection WebSocket, parallel task sessions with streaming cards, media in/out, cloud docs) + restart recovery notification (ops).
多 agent 团队协作:并行调试、并行功能开发、任务协调、沟通协议、团队构成。受 wshobson/agents(38k★ MIT)启发。
Use OpenCode models inside DeepSeek Harness without a separate provider API key.
Readable thinking display for DeepSeek Harness: live per-paragraph summaries of the reasoning stream, rendered in the chat history where the 'Think' row is, with a user-chosen summary model (thinking disabled on summary requests).
Two-level fold for DeepSeek Harness chat: collapse think+tools as a process, then fold tool rows separately. Official cards stay intact. 聊天主视图两层折叠:过程一层、工具一层,不替换官方卡片。