Plugins

Desk-to-desk tickets between sessions on one machine. A session says which desk it belongs to by putting a number prefix in its title; a ticket names one desk and is taken by the first session of that desk to start working, so the desk is the address and a ticket is not delivered to a session that is just passing by.

Prompt optimizer for the composer: a button in the tool row rewrites your draft through a model channel you already configured, streaming into a panel that floats above the composer without taking up layout. Your draft changes only when you click Accept — and Accept is refused while the answer is still streaming, so a half-finished reply can never be written back. It asks first when the draft carries @ references or attachments, which a whole-draft replace would flatten into plain text. No global keyboard shortcuts, and no API key ever reaches the browser: the host half makes the call.

A project-scoped memory plugin for DSH. Knowledge authoring and the plugin's own prompts are open to editing, so users can redesign the library structure and how the model is driven without changing a line of code. By default the library is built on one core idea: a project document you design yourself (a handoff by default) is the project entity; the plugin detects which project a session belongs to, binds captured knowledge to that project, and promotes what generalizes into theme and global layers. The most natural logic there is does the pre-filtering of knowledge, with controllable compression and archiving of context.

Model attribution panel for the dsh web chat — local, zero-network structural fingerprinting (template token leaks, glitch canaries, tokenizer usage-vector matching) that scores which of 29 AI vendors a session's responses most resemble, with a per-claim evidence ledger and a 12-probe active battery.

A chemistry workbench panel in the DSH web UI. Draw structures on a Ketcher canvas (a script downloads the build; it renders locally in the browser with no backend), then compute on the current molecule with RDKit without calling a model — formula, molecular weight, exact mass, logP, TPSA, hydrogen-bond donors and acceptors, ring counts, Fsp3, stereocentres, Lipinski and Veber rule-by-rule checks, QED, PAINS/BRENK structural alerts and the Murcko scaffold, plus SMARTS substructure matching, fingerprint similarity, format conversion, standardisation and batch clean-up. A retrosynthesis tab talks to a pluggable backend (Retro* by default) over a single JSON-in/JSON-out process, and states what is missing when the backend is not installed instead of inventing a route. The calculations run in this repository's Python package (rdkit), which is installed with pip separately.

On-demand skill tools for DeepSeek Harness: registers skill_search / skill_load / skill_ref so the model searches a staged skill library first and loads a skill body only when it needs it, keeping the library out of the resident session catalog and out of every turn's context; adds a Skills tab to the conversation view ring that reads back the whole session and lists which skills were actually loaded, and in which turn. Zero dependencies, and the records render client-side only, never entering the model's context.

Security-audit facts AI coding agents cannot compute, plus a verdict gate: euthyna history attributes deleted lines to commits and flags those from security fixes (--origins finds the first introducer), euthyna coverage reports which changed symbols no test ever invoked, and euthyna gate checks each finding against six gates, downgrading any without evidence to an observation.

Adds an Explain button to the approval card: one model call returns exactly three lines reading the pending operation, covering what it does, which files or system state it reads, writes, modifies or executes, and a risky/attention/safe verdict with a one-sentence justification. It also renders the detail area for `write` and `edit` calls, which the shipped renderer leaves blank because it only reads a `command` field.

Filesystem permissions for shell calls, granted per path and capability rather than per command name. The command line is parsed for the read, write, create, delete and execute effects it needs, a capability it lacks raises the approval card instead of a flat refusal, and the same rules are compiled into the macOS Seatbelt profile the process, its children and the code its arguments never showed all run under. An Approvals tab in the conversation reads the audit log back and shows which rule, automatic reviewer or person answered each call.

Plugin-author upgrade skill for DeepSeek Harness: one package carrying a corridor index, which detects the caller peer band and routes to the closed corridor card that matches it (0.1.3-alpha.1 -> 0.1.5-rc.1 as legs A+B, 0.1.5-rc.2 -> 0.1.6-alpha.2 as leg C), plus a zero-dependency seam scanner shipped as a bundle skill and an npx CLI.

Per-workspace MCP scoping for the DeepSeek Harness — decide which MCP servers a session may see from the directory it was opened in.

Reports DeepSeek Harness session activity to a self-hosted cc-monitor server, which surfaces Claude Code and DSH sessions as color-coded state cards in one web dashboard and pings your browser and Android app when a session needs attention.

China-ready voice input for the composer. Requires a local Python bridge (pip install dashscope websockets, run bridge/voice-bridge.py) — the plugin alone does not work. Browser mic streams 16 kHz PCM to the bridge, which runs Alibaba Cloud DashScope ASR (paraformer-realtime-v2); interim text fills the draft at the cursor, silence auto-stop, optional auto-send.

A green interrupt button beside the send box: silently stops a running agent, which then summarizes its progress and asks for your new requirements, with a customizable interrupt prompt.

DeepSeek Harness (DSH) plugin: a "Thinking Language" selector under Settings → General that tells the model which language to reason in (chain of thought) via a system-prompt section. Zero extra calls, zero latency, 12 languages.

Recursive delegation for dsh agent trees: a preset plus Host limits bound what each subagent may do, subagents work in separate _agents/ workspaces, and the parent decides which results to integrate.

Keeps a DeepSeek Harness run going on its own: continues a turn that ended abnormally, and continues a turn logged as completed while the model tool call was emitted as plain text instead, which leaves the agent silently idle. Revives a model request that stops streaming. A wait on an approval or on a question is logged and notified, never answered.

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.

One settings page for every MCP server on your machine. It reads the real config file of 17 other AI tools (Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, OpenCode, Trae, Trae CN, Qoder, Gemini CLI, Qwen, Zed, Roo Code, Copilot CLI, Kiro, CodeBuddy, Continue) plus DeepSeek Harness itself, groups servers by the file that declares them, probes each one over JSON-RPC to see whether it starts and which tools it exposes, and gives every server a switch that mounts or unmounts it in DeepSeek Harness without editing the other tools' own files.

Read-only assessment of whether a Dameng DM8 database can be recovered and to which point (LSN or time), across 16 recovery scenarios, with an HTML report. Needs Java 8+ and a Dameng JDBC driver jar (DmJdbcDriver8.jar) that you provide yourself.

Adds job-role prompts to the DSH composer: type - at the start of the input to pick a role, which becomes a reference chip in the draft and is then injected as a system prompt section on every turn of that conversation, so the instruction keeps applying without showing up in the transcript. Roles are added and edited on a Settings page and stored in ~/.dsh/dsh-role-prompts.md.

Predicts your next message after each completed turn and shows it as ghost text in the composer: press → to accept, or keep typing and finish it word by word with Ctrl+→, Esc to dismiss. The prediction model follows the session by default or can be picked in the settings card, which also has an on/off switch and a speed test.

Keeps quotes out of the message body: quoting any passage from an assistant turn opens an editable quote tray above the composer, the quotes are injected into the next turn as system-prompt context (attributed with source and time), and an expandable quote row under that turn shows exactly which quotes were used. Quotes persist across page reloads and are claimed per turn; the host and client halves hand-shake on version and report a compatibility verdict.

Publishes whichever model the profile has selected in chat or on the Models page as an OpenAI-compatible HTTP endpoint: /v1/models, /v1/chat/completions (streaming SSE and non-streaming), /health. A request may name any routable model; one that omits `model` gets the current chat selection, so switching model in chat switches the endpoint. Switched on, it can also lend the model the Harness tools switched on individually, run those calls through the Harness tool runtime and feed results back up to a round limit. The listener is the plugin's own, so 127.0.0.1, 0.0.0.0 and a NAT-forwarded port all behave the same with no change to the request path. Settings cover bind address, port, API key with a generator, fallback provider and model, reasoning effort, CORS origin and timeout.