Plugins

Model channel auto-enricher for DeepSeek Harness: eliminates manually configuring contextWindow, maxTokens, \[text, image] multimodal input, and {low, high, max} reasoning effort tiers model by model after adding a provider channel, with live auto-enrichment on save, a Settings UI page, and timestamped backups.

Overhead conversation background with five whale species formed from custom text or ASCII, shaded cells, frosted content cards, input-box avoidance, independent dives, spouts, ripples, wind, and cloud shadows.

Per-turn tool cost line in the chat stream: after each turn a collapsible row under the reply lists which tools ran, how many times, and how many tokens each tool injected into the context (input plus output), expandable to per-tool detail.

Learning Mode toggle in the input toolbar: injects a teaching prompt into the current session so the assistant introduces a topic, teaches it, tests it, and corrects mistakes, rendering graded quizzes through GenUI. Requires the GenUI plugin (@changfenhuang/dsh-genui); without it the teaching text still works but quizzes, sliders and charts degrade to plain text.

Cuts session input tokens three ways, each with its own switch in a panel under the composer: lossless structure-aware compression of tool results on `tools/post-execute`, then a per-tool head/tail elision that always names a recovery path through the shared spill store; tool-catalogue visibility per agent on `agent/created`; and a per-workspace memory document that recall matches against each new message and injects through `agent.inject`, so the transcript is not re-read. Measured 41.76% fewer prompt tokens over 149 recorded requests, with a bundled script that re-measures any profile.

Persists unsent composer attachments per session, so images and files pasted into the input box survive page switches, conversation switches and app restarts: the bytes are mirrored in IndexedDB and handed back through the official createDrafts/addAttachments channel, and are dropped as soon as the message is sent or you remove them.

Claude Code-style prompt history, double-Esc picker, selection quote/copy & right-click paste.

DSH Web plugin that injects an SVG (feTurbulence + feDisplacementMap) refraction filter and applies it through backdrop-filter to panels, bubbles and the input card, and bundles the orca-scene skin with theme tokens plus light and dark backgrounds.

Maintained fork of dsh-approval-gate. A neutral operation whose confirmation count has reached the threshold auto-approves when the judge is unavailable, instead of prompting a human again; approver-facing explanations are generated in Chinese from the real sandbox mode, command and paths. Also carries deterministic hard-deny for credential exfiltration and system-path destruction, judge-input redaction, a judge model candidate chain, fingerprint-scoped allow rules, and an approval view with unified diff and one-click revert.

A local-first personal state record and decision engine for dsh: keeps every input byte-for-byte with line-level provenance, argues against its own conclusions by generating defeaters first, ages each stored fact with its own TTL, and reports where a decision rests on nothing at all.

Model capability editor: edit each model reasoning-effort levels and image-input (multimodal) declarations inline in the official model page, written back to settings.yaml as a whole group, with a floating entry as fallback when the anchor breaks.

Windows computer-use plugin for AI agents with reusable window targets, UI Automation, window-bound screenshots, background-first verified input, isolated Chromium automation, MCP, CLI, and installable skills.

Voice call assistant for the DSH desktop app and the Web GUI (the same plugin, only the profile name differs): a draggable floating call ball, browser-side VAD that sends each utterance after a pause, speech recognition through the local SenseVoice model already downloaded by the official voice-input plugin (no second model), FunASR HTTP or streaming, or a cloud-compatible API, MiniMax / MiMo-V2.5-TTS (director mode and voice clone) / OpenAI-compatible TTS replies, an optional wake-word mode with auto-sleep, and task dispatch to separate sessions with progress, stop, and spoken completion reports.

Scheduled tasks panel for managing DSH agent and system crontabs with natural language input and execution logs.

Skill marketplace with aggregated search across SkillHub and ClawHub, in-app README preview, one-click install and uninstall, and a chat-input skill picker.

Recall previously sent messages in the input box with Up/Down, and archive a workspace session from its right-click menu.

Volcengine Ark Coding Plan LLM provider for DeepSeek Harness: provider-scoped PDF and image input, with Responses API streaming support.

OpenAI-completions-compatible adapter for custom gateways (vLLM / LM Studio / self-hosted proxies): always role:"system", thinking driven by the model config, Qwen-style response split, and vision-model image input (single / multiple).

Input prediction for the DSH composer: your next prompt suggested as ghost text after each turn, accepted with Tab. (autocomplete / input prediction / prompt suggestion)

Draggable ring gauge showing live model output speed (tok/s) with an expandable per-step detail panel (exact speed, TTFT, cache-hit rate, cache read/write, reasoning and input tokens, model, cumulative tokens).

Adds a sidebar tab that folds a session into user turns (input, tool steps, output) with one-click jump back into the conversation.

Prompt-cache diagnostics dashboard showing provider-reported token usage, adjacent logical-input changes, and Call TTFT.

String-matches tool-call input arguments, blocks dangerous tool calls (deny) or allows them with an injected warning (warn), and ships a full rules-management panel.

Turn a phone camera into a live viewfinder and photo input for your dsh session, over LAN or USB.