Plugins

Model-driven context-window rollover bundle for the DeepSeek Harness: fresh working context, durable model-managed notes, a small verbatim recent tail, and targeted history recovery — no summarization.

Fills in what the official pi-ai adapter leaves out on the Models page: gateway reachability probing with latency, one-click pull of the gateway model catalog with checkbox write-back into settings, DeepSeek balance and provider quota shown in place, the protocol and endpoint each provider is actually wired to, and a model dropdown that lists only what works right now — a dead key, an empty balance or an exhausted quota is hidden until you fix it, and comes back on its own.

Embeds external AI chat sites in a DSH side panel through the same-origin reverse proxy with per-site DOM adapters, and adds a composer button that either rewrites the draft with the host model or runs a two-round loop: compress the conversation, optimize, send to the embedded site, capture the reply, then finalize it into the composer.

Groups the DSH model picker by provider: the menu lists sources with their model counts, and hovering a source slides that provider's models out in a side panel (default); in panel mode a click opens the same panel, and double-clicking a source title cycles hover -> panel -> inline (expand below the title). The active model stays visible at a glance. Pure DOM enhancement with zero intrusion - the panel is plugin-owned DOM mounted on body and selection happens by programmatically clicking the real menu buttons, so no React-managed node is moved or removed, keeping the host rendering intact and coexisting with other UI plugins. Styling follows the host by reusing the official --dsw-* theme tokens, with an enforced opaque background, sufficient contrast, and one shared set of sizing and colour tokens for both menu levels. No network access and no telemetry. Built for DSH kernel 2.0.15-next.

Automatic approval and review for tool calls that need approval: keyword buckets match red lines without context, then a judge model classifies what was requested and sends it to allow, reject or a human according to the matching row and risk level; a judgment that never ran always goes to a human, and rejections tell the model why.

Visualize the plugin and model call flow as a 3D office with six fixed workstations, thought bubbles and a delivery sheet.

Workspace-wide sticky notes on an infinite canvas: notes are plain Markdown files, created manually, captured from conversations, or distilled by the AI, and can be handed back as context; ships 12 model tools and 5 bundled skills for creating, tagging, arranging, comparing and merging notes.

LLM provider adapter that routes model calls through any ACP (Agent Client Protocol) server — Claude Code, Codex, Gemini CLI, Devin and more — with a registry browser, per-server settings UI, lazy authentication, and dynamic model discovery.

Replaces the native DSH reasoning-level option list in place with a draggable Off, Low, High, and Max slider while preserving the official model-selection flow.

Notched reasoning-effort slider grouped with the model picker in the composer, with theme variables and English, French and Chinese labels.

Adds selection annotations, in-place translation, sidebar subagent questions and default-browser link opening to DSH. The selected text rides along with your next message, the source keeps its highlight and ordinal mark for jumping back, the menu opens on selection or from the right-click menu, and the translation model can be picked separately.

Remote-control bridge: pair a running dsh web/desktop client with the Happy mobile app by QR code, then remote-control the same harness sessions from the phone (send messages, read replies, approve tool calls, switch models)

Resumes an agent turn that ended after narrating its next action without calling a tool: the guard replays the turn log on agent/turn-stopping, asks a judge model one true/false question, and steers the same turn to run one more step.

3D printing & additive manufacturing plugin for DeepSeek Harness: parametric STL generation, calibration G-code, STL/G-code analysis, toolpath rendering and PrusaSlicer bridge, plus 5 skills and a concise persona optimized for local models.

Expert manager: manage ntd-format experts and expert teams (plugin.json + Agent MD + skill sets), browse and install from a built-in market of 50+ experts, and run tasks in an expert persona via /expert-<name> without consuming model directory tokens.

Multi-model orchestrator: the main brain decomposes complex tasks and dispatches subtasks to GLM / Kimi / Qwen sub-agents by each model's strengths, with per-model token usage.

Irreversible secret-scrubbing guard: rewrites access keys, bearer tokens, and private key blocks into `[REDACTED:<category>]` placeholders before they reach the session log and the model.

Cut token cost in 4 layers without cutting model intelligence: handoff, rtk, headroom, caveman token-saving skills — each also available as a standalone installable plugin.

LLM request retry plugin for DeepSeek Harness: when a model request fails with a configured HTTP status code, machine code, or provider field=value match, it sleeps for the per-rule delay and re-issues the request automatically, so subscription quota and rate-limit errors (429 / insufficient_quota) do not interrupt a run.

Claude Code integration plugin for DeepSeek Harness, providing native Claude conversations powered by the Claude Agent SDK, with model selection, approvals, questions, tools, and session continuation.

Command Code Provider API chat for DSH with a Host-held API key, discovered model catalog, effort defaults, an optional zero-data-retention header, quota meters, and per-model OpenAI or Anthropic protocol selection.

AI audio generation for the DeepSeek Harness web GUI — multi-vendor TTS, music, sound effects and voice design with a sidebar panel, model comparison, resource library and Agent tools.

Group installed skills and manage MCP servers from a left-panel UI: filter the model skill catalog by enabled groups (union-deduped) and toggle MCP servers independently.

Read-only insight panel for one profile: five views over the same data — config layers, plugin tree, services, tool names and models — each entry showing which layer inserted or disabled it and whether it is running.