Plugin marketplace

Long-term memory for DeepSeek Harness kept as plain markdown, one fact per file, readable in any editor. The `MEMORY.md` index has a hard 200-line / 25 KB cap enforced by `ctx.tools.guard()`: a write that would grow an over-cap index is denied, while a rewrite that shrinks it always passes, so an oversized index can be compacted step by step. The plugin never writes memories on its own — what is kept is decided by you and the model. The same store is also used by Claude Code, Codex, Kiro and OpenClaw.

Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified: a compatibility layer implementing Pi's public extension ABI on top of DSH's native services, so Pi's hundreds of mature packages (web search, memory, code navigation, subagents, vision…) install straight into DSH — `dsh plugin add pi2dsh` once, then any Pi plugin from npm as published.

Local Git-backed long-term memory for DeepSeek Harness with cited context, usage-aware read tools, safe sync, previews, and rollback.

A local creator workbench on DSH: from topic and script through screen-recording projects to subtitles, covers and publish status — "one video, one folder." Manuscripts and artifacts stay plain files any editor or AI tool can read, while workbench state (project bindings, publish marks, synced playback stats) lives in a separate overlay instead of a closed content database; Screen Studio, subtitles, covers, WeChat-article and publishing capabilities are all optional installs.

The reasoning selector rebuilt as a slide rheostat: a 0–30 "Liang intensity" dial reading reasoning.efforts live, syncing 24 character portraits, background and UI colors as you drag, snapping to the nearest level on release

Unofficial Command Code LLM provider: registers a `commandcode` route with a live model catalog and reasoning-effort support.

Local-first telemetry collector for AI coding agents — unified OpenTelemetry events for Claude Code, Codex, Cursor and more. Token usage, cost, traces and security audit, exported anywhere.

A find-plugins skill for DSH: say 'is there a plugin that…' and the agent searches the whole dsh-plugin topic on GitHub, waits for your call, then installs and verifies per the repo's current claims (full flow in skills/find-plugins/).

Persistent multi-model squads configured in Settings and selected in the Composer, with per-member model/tool policies; the main Agent dynamically plans each send as a bounded DAG with bounded reviewer/repair loops, while Run Center tracks retries and per-member tokens from official Harness usage events.

A multi-device remote access solution built on DSH's plugin mechanism: a secure, low-latency, end-to-end-encrypted, P2P-first network to reach your home DeepSeek Harness from PC, Android phone, tablet or browser — with companion Windows/macOS/Linux desktop clients, an Android app and a web client.

Multi-device remote access for DeepSeek Harness: continue an active session from your phone, tablet, browser, or another computer over an end-to-end encrypted channel (Noise IK + adaptive relay/WebRTC transport), with device authorization and ApiProxy-only remote capabilities — no shell, remote desktop, or file access.

OpenPencil design preview and editing plugin.

Anime search plugin for DSH: multi-source in-chat search, Bangumi rating cards with details, copy magnet links

金融与产业研究的一屏式实时行情大屏（Market Research Cockpit）：A股/港股/美股指数、大宗商品、美债收益率、板块热力、主力资金流、724 快讯与产业链自选股，附 AI 大模型解读；免 API Key、免登录即开即用，附在线 demo。

Moves an existing DSH session to another agent preset through a previewable five-section handoff, preserving the source session and either pausing the target for confirmation or continuing immediately.

Export DSH conversations as Smartisan Notes-style PNGs, or create and update Markdown notes in a configured account-scoped workspace.

Export DSH conversations as Smartisan Notes-style PNGs, or create and update Markdown notes in a configured account-scoped workspace.

GitHub-topic-driven plugin & skill marketplace: a Settings page that browses the auto-collected registry (the whole dsh-plugin topic plus the skills index, CI-refreshed every 2 hours) with one-click install, type detection, install-script and host-shadow-dependency safety confirmations, env-key management, and the STANDARD.md recognition spec.

Free, keyless web search for DSH: 7 engines (DuckDuckGo/Bing/SearXNG free + Exa/Perplexity/DeepSeek paid), auto-failover, settings-page UI with API key inputs and official links, web_fetch, and an engine test tool.

Turn the chat screen into a visual novel: 16:9 stage, whale-girl maid sprite, ornate dialogue box, typewriter lines — plus a drag-and-drop scene editor that syncs straight back into play mode

A BepInEx-style runtime mod injector: inject any local plugin package into a running dsh web — no patches, no package.json, no bundle lists, no restart; full effect (host tool + client UI) on injection.

Tracks DeepSeek token usage, per-call and session costs, and account balance with cache-aware charge animations in the DSH Web UI.

AGI long-term memory base: multi-agent spatiotemporal memory graph, self-evolving knowledge flywheel, self-cognition, and auditable trust guardrails.

The missing middle ground for DSH permissions: routine project work runs directly inside the official workspace-write sandbox, while semantic risks outside it are classified by the current DSH model against your instructions — genuine ambiguity asks once, and destructive access to critical paths is denied before execution; an alternative to both interruption-prone restricted modes and no-approval Full access.