Plugins

A `describe_image` vision tool for text-only models: images (local path, URL, attachment) go to a configurable OpenAI-compatible vision endpoint and only the returned text enters the session.

A fixed-corner whale widget for the DSH web GUI — balance, today's usage and per-turn cost with peak/off-peak pricing, editable balance-alert and daily-budget bubbles, a module-based custom bubble queue with A/B weighted choices and random lines or images, 30+ vendor templates (OpenAI, OpenRouter, Kimi, SiliconFlow, Ark, Zhipu, MiniMax and more) with per-model balance and subscription quota, plus task-end sound, imported audio, custom roles and a resource manager. Local-only, no telemetry.

Chat recovery for DSH: edit the last finished user message and explicitly retry failed turns, both via session-fork branches that leave the original conversation untouched.

Runs project OMA checks before and after tool execution and when a turn tries to finish. Installs its OMA CLI dependency and initializes unconfigured projects on first use. Requires Node.js 22+; initial setup needs network access and project write access.

A purely plugin-based implementation that brings "cross-session long-term memory + background self-evolution" capabilities to DeepSeek Harness: five-track memory · git branch awareness · in-turn self-review · skill self-evolution and skill manager · four-track to-do · COI scheduling · session broadcast · session search · prompt manager · temporary info notes—zero core modifications, zero runtime dependencies, ready to use upon installation and completely removable upon uninstallation.

Connects dsh to a local SandBase Harness runtime over stdio MCP for managing agents and sessions, streaming turns, inspecting artifacts and cancelling work.

Turns a video (a local file, or a Bilibili/yt-dlp link) into one addressable storyboard grid — numbered cells with exact timestamps, per-cell transcript and an audio envelope on the same timeline — and video_zoom re-reads any window at higher resolution, so a text-only model can look at the video instead of reading a summary.

Per-session long-term memory that fades over time, recalls by evidence score and injects a few entries into the system prompt each turn.

ThoughtDAG for DeepSeek Harness: turn the current DSH session into an editable thought graph on an infinite canvas.

Rebuilds the DeepSeek Harness web sidebar, workspace tree, search, and turn navigation in a Codex-style layout.

Gives the companion a mood judged each turn and shown as a short context note, with a panel for mood, emotion dimensions and timeline.

Second-model auto-review on the approval answerer chain: a read-only reviewer subagent returns structured allow/deny verdicts with reasons, fail-closed by default.

Full-screen 550C boot intro for DSH: a 4-second simple cut and a 16-second rewrite sequence, skippable by click or Esc, with four phosphor palettes. The host half injects the opening frame while the document is still parsing, so the built-in loading card never shows; the desktop caption strip turns transparent, so those window buttons float on the animation.

Rotates the dsh turn-status label through 1059 bilingual phrases with typewriter output, an animated rainbow gradient, danmaku, 12 toggleable theme packs, live placeholders and tab-title rotation. Two star packs ship disabled by default; the stargazer pack carries one phrase per stargazer, refreshed weekly by a workflow.

Turn your phone into the control center for an AI coding assistant CLI session (Claude or Codex) running on your computer.

Reports which skills each turn invoked (loaded by the model, typed as /name, or none) and manages the local skill directory: install from a pasted repository, folder, SKILL.md or zip address, set a Chinese display name in the skill meta.yaml, and remove a skill while keeping a backup.

Study workspace for DeepSeek Harness: imports your own PDF, DOCX, PPTX and Markdown materials plus lecture recordings, turns them into source-cited flashcards, quizzes and open questions, and schedules spaced review with FSRS-6 in a built-in web panel. Imported recordings are transcribed with your own SiliconFlow, Groq or Gemini key; live classes can instead use an on-device SenseVoice engine that needs no key (about 239 MB on first install); requires DSH 0.2.0-rc.2 or later but below 0.3.

Google Antigravity (agy CLI) models for DSH — streaming chat with Gemini/Claude/GPT-OSS subscriptions, native tool cards, thinking turns, and in-GUI Google OAuth login.

Rewind conversation and workspace state, powered by a persistent Change Ledger.

Browse Claude Code, DSH and Codex session logs by working directory, with per-turn context composition, token usage, compaction markers and cost estimates.

边做边压 Live Compact. Compress the finished turn in place and keep the prefix cache.

Code-first React and TypeScript task apps rendered inline, in Canvas, full screen, or on localhost, with interaction state available to later agent turns and approval-gated MCP and API access.

Read-only session work-path observer: fold turns into stage, run, duration and tok/s summaries, then expand parallel branches, tool results and vendor-visible reasoning.

Turns explicit DSH session experience into reviewable native Skill drafts that are saved only after user approval.