Tools & Capabilities

Opens the official chat.deepseek.com page in the DSH main panel through the desktop app's native browser view, so the login survives restarts, without an iframe or request proxying. A fork of JaMoLpE88/dsh-desktop-chat with the changes listed in its NOTICE. The login token is kept in plaintext in session.json.

Runs local CLI agents (Claude Code, Codex, Antigravity and Pi by default, auto-detected) as panels inside DSH, reusing the official subprocess.spawnTerminal with an embedded xterm.js. Names, commands, colours and order are configurable. Requires DSH 0.2.2-rc.1 or newer. The CLI agents must already be installed and run with your own user permissions.

Reads bilibili video as structured text for AI: metadata, comments, danmaku, subtitles, local ASR, frame OCR. Local processing (no cloud upload), decoy-subtitle detection with frame check, instant re-reads via cache, AI-ready manifest + summary. Frame OCR needs ffmpeg; local ASR/OCR need Python 3.10+ venvs (plus the VC++ runtime on Windows); AI subtitles need a Bilibili login.

OMK — Observe. Measure. Know. Evidence-backed knowledge changes for AI applications.

Bilibili video analysis: metadata, transcript (ASR fallback via Bijian/sherpa-onnx/whisper.cpp), comments, danmaku, and sharp keyframes with optional local vision descriptions.

Linkr Bee (BLE/UART wireless serial bridge) as a DeepSeek Harness plugin: a live xterm console in the right column, driven by keystrokes, plus agent tools that run commands on the target board through bee-cli.

Reads ComfyUI PNG metadata in a model-facing comfy_png_inspect tool: point it at one PNG or a whole output folder and it reports, per image, the LoRA execution chain with every strength value, sampler settings, base model, CLIP, VAE, latent size, both prompts and bypassed nodes.

Union web search for DSH (consolidated successor of DDDMUC/dsh-free-search, which it credits): 28 engines across three tiers — keyless free (Bing/DuckDuckGo/AnySearch/SearXNG multi-instance union, plus Exa/Tavily/Keenable/Firecrawl/Parallel keyless modes), paid-with-key (Zhihu/Perplexity/Serply/DeepSeek official/You.com/Baidu/Kimi/Aliyun/Doubao/Zhipu/Bocha/Metaso), and model built-in search (OpenAI/Gemini/Claude/Grok, explicit opt-in only) — with auto smart routing, per-engine enable/disable & fallback priority, parallel top-3 merge, multi_search, failure-aware fallback, a toggle to block the system built-in search, time filtering, platform search, web_fetch and a web settings UI. Experimental free advance_fetch: whole-page GFM-Markdown or CSS-selector/fields structured extraction with no API key and no per-call billing; fetch identity configurable (custom User-Agent, Referer, per-domain cookies attached only on domain match, never leaked across cross-domain redirects). Hosts 0.1.0-rc.2 through 0.2.1-rc.2 (17-tag exact engines.dsh enum; the rc.2 search-provider takeover goes through volatile config pinning after capability probing, never version branches). Paid engines need your own API keys (optional; free engines work keyless). npm: @drscrewdriver/dsh-union-search and bare dsh-union-search.

Map your Zotero library into DSH: let AI operate Zotero for you and help you read papers. Browse collections and tags in the sidebar, read PDFs with highlights in place, and cite excerpts, annotations, notes or whole items into the chat. The model works from the source text for reviews, comparisons and data extraction, and can also search and organize the library, import papers and attach translations back to items; first use needs the Zotero MCP plugin and the local API enabled.

Aggregated multi-upstream web search for DeepSeek Harness: replaces native web_search with model-native grounding and agent search APIs, source attribution, and rich citation cards.

Business knowledge-network context for DeepSeek Harness: bind a session to an OpenBKN knowledge network, query its objects and metrics through managed MCP tools, and open trace-backed provenance (execution facts and a business graph) for each answer. Ships with the OpenBKN Runtime; the provenance views need an OpenBKN enterprise license.

Multi-platform research and deep search pipeline: harvest_scout for parallel discovery across GitHub, Web, Twitter, Reddit, Xiaohongshu, Bilibili, and YouTube, harvest_deep_research for automated comprehensive research reports, plus extract, cross-source verification, and credibility auditing.

Computer use through an external USB HID device: name a target in plain language, a vision model locates it in a cropped window screenshot, and a Raspberry Pi Pico 2 moves the real mouse and types on the real keyboard. 7 tools. Windows has window listing, per-window cropping and automatic window raising; macOS needs pyobjc and has neither; Linux is unsupported.

Renders mermaid/mmd code blocks in the conversation as chart cards with preview/code toggle; the engine is inlined so it works fully offline with no CDN (npm: `dsh-mermaid-render`).

Deterministic context compression for dsh — zero LLM calls, rule-based pruning that preserves paths, commands, error traces, and open todos verbatim. Compresses ~80k tokens to a few hundred.

Ten video tools (probe/cut/concat/encode/subtitle/extract/gif/frames/adjust/health) on the official subprocess service: probe summaries, batch frame extraction, speed/volume/mute/rotate adjust, shell-free argv, zero runtime dependencies.

Liang Saint Slider as the DeepSeek Harness model + thinking-effort selector.

Load only basic DeepSeek Harness tools by default and activate other tools on demand.

Research Agent skills, literature search with citation tracking, document-to-markdown knowledge base builder, and in-depth reading with browser workbench for DeepSeek Harness.

General-purpose LLVM static-analysis plugin for DeepSeek Harness: builds analysable bitcode and runs kernel-analyzer (KAMain) reachability analysis behind ctx.kanalyzer

DSH plugin bridging agents to Cheat Engine for dynamic memory debugging: scanning, pointer analysis, breakpoints, and game memory hacking.

Low-cost video understanding tool: a video_understand tool turns a Bilibili link / BV / local video into an AVIS info layer (ASR + scene structure + object tracks + YOLO labels) and returns summary + Q&A. Question-driven dynamic routing across layers (L0 ASR / L1 object tracks / L2 key-frame VLM), semantic-layer reuse for repeat questions, per-question budget cap. Python engine: core layer needs faster-whisper / opencv / yt-dlp (~200-300MB); the optional semantic layer adds ~2GB of torch / transformers / ultralytics. A bundled doctor --fix sets up the venv and installs both.

Hermes-style tool search & slimming: progressive disclosure, semantic search, describe, and call long-tail tools on demand while core tools stay eager.

Reverse-engineering preset for DeepSeek Harness: domain skills live inside one "reverse mode" and the model routes by target type (web/JS, Android, iOS). The JS domain ships three skills (AST deobfuscation, controlled iv8 execution, JSVMP recovery) and mounts the js-reverse-mcp browser forensics server as mcp__js-reverse__* tools.