Plugins

Pacote opcional somente-Skill do DeepSeek Harness para a skill de diagramas de arquitetura Archify.

Swaps the filesystem and bash providers for a mirage virtual workspace: file tools and shell commands run over mounted resources (RAM, S3, Redis, Slack, Gmail, Notion, Postgres) instead of the host disk, with per-mount read/write/exec modes, per-command sandbox routing (monty, pyodide, quickjs in process; docker, e2b, daytona remote), and installed CLIs (git, gh, slack, linear, ntn, gws, or one you register) as head words in the virtual terminal.

Ponte de visão para modelos somente-texto: cole uma imagem e receba evidências estruturadas em JSON (OCR, layout, semântica).

DeepSeek Harness tool plugin that exposes BrowserSkill (bsk) browser automation to the model

Runs local security audits of AI API relays and LLM proxies from DeepSeek Harness, producing Markdown reports for prompt injection, model substitution signals, tool-call rewriting, error leakage, stream integrity, and profile-gated Web3 risks.

Tarefas de visão para modelos somente-texto: perguntas e respostas sobre imagens com reconhecimento de intenção, OCR de capturas de tela longas, reprodução de UI, grounding e diff de pixels.

Runtime de agente open source compatível com CMA para qualquer modelo, com ferramentas MCP, sessões em sandbox, auditoria, replay e um console local. Inclui um pacote nativo para o DeepSeek Harness via MCP stdio.

Componentes de UI interativos renderizados diretamente nas respostas: layout, gráficos, formulários, quizzes, mermaid, cenas 3D e um loop de eventos de ação de volta para o modelo.

UI generativa dentro da conversa: o modelo renderiza cartões HTML interativos diretamente no fluxo do chat, com preview em streaming e renderização em sandbox.

Sidebar usage panel that attributes tokens to the relay site that served each request, read from your existing provider config: today/month/all-time totals, per-site and per-model breakdowns, a year activity heatmap, and New API / Sub2API / DeepSeek balances.

Adaptive reasoning slider skin that maps each model's available reasoning efforts onto a 0–30 visual intensity scale, with synchronized portraits, background, and interface colors.

DSH适用的Codex风格的思考强度滑块，以及大肥鱼跑步滑块。Codex-style model and reasoning-effort slider for DeepSeek Harness

Mapa de calor de uso de tokens, detalhamento por modelo e saldo da conta DeepSeek para a GUI web do DeepSeek Harness (dsh web).

Plugin não oficial de provedor de LLM do DeepSeek Harness para o Command Code: catálogo de modelos ao vivo, suporte a nível de esforço de raciocínio, card na página de Modelos. Portado do pi-commandcode-provider (MIT).

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.

DeepSeek Harness (dsh) bundle wiring SkillRoute's skill router into your agents.

Let Your Ideas Flow — Tydora is a modern desktop Markdown editor combining WYSIWYG editing, bidirectional links, mind maps, and an infinite canvas — empowering deep thinking and effortless expression.

Exponha ao modelo as ferramentas de parsing de documentos do MineRU.

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.

Word alternatives and underscore-gated fill-ins in the composer. End a line with _ and it is filled; misspellings are flagged as you type. Routes through the model dsh is already configured with, so it needs no API key.

Canal de bot de mensagens Lark/Feishu para o DeepSeek Harness: conversas acionam agentes, e respostas e aprovações retornam como mensagens e cards

Silky streaming reveal for the Web UI: text appears at the model's arrival rate, new lines glide in, no flicker; follow stays with the user and prefers-reduced-motion is respected.

Modo Auto para o DeepSeek Harness (dsh): uma opção Auto no seletor de Permissões cujos prompts de aprovação são respondidos pela política do OpenGuardrails em vez de um humano — além do motor de proteção OGR completo por trás. Sem alterações no núcleo.