Plugins

Página Skill Manager no painel de definições do DSH: navegue por skills de sistema/utilizador/workspace/predefinidas, expanda uma skill para ver a respetiva árvore de ficheiros, veja e edite ficheiros, importe skills de um zip e exporte-as ou elimine-as (skills de sistema são só de leitura).

Gerenciador de assistentes no estilo Hana para DSH: gerenciamento visual de predefinições (cartões, personalidades yuan, recorte de avatar) e uma pilha completa de memória Hana (resumos contínuos, compilação em camadas, armazenamento de fatos SQLite, memória profunda).

Atalhos de sessão para a interface web do DSH: alterne sessões como abas do navegador (Alt+1-9 posicional, slots fixados, anterior/próxima, nova/arquivar/renomear), troca de modelo via teclado com Alt+M, envio alternativo, repasse de foco com Esc e Alt+W para fechar ou restaurar um espaço de trabalho dsh-split-screen — tudo reatribuível com predefinições para Windows/macOS. Verificado no DSH 0.1.1 e 0.1.5.

Dock de cota do plano de tokens MiniMax para a interface web: um chip em forma de gota com a % de uso de 5 h na área de entrada da conversa; clique abre um painel detalhado listando o uso de 5 h e semanal de cada modelo e a contagem regressiva de reinício, ocultado automaticamente quando o modelo padrão ativo não é um modelo MiniMax.

Substituições por chamada de modelo, provedor, persona e toolFilter para delegação a subagentes, com referências @preset: e ids compostos de provedor/modelo

Scanner local limitado e que preserva privacidade para metadados de Codex, Claude Code, WorkBuddy e CodeBuddy: agrega atomicamente evidências chaveadas e aconselha deterministicamente presets DSH embutidos e capacidades opcionais — sem retenção de conteúdo, sem acesso à rede e sem mutação de presets.

Custom modes and custom prompts for DeepSeek Harness (dsh): edit a mode's system prompt on the settings page (it takes effect on the next model step), choose its base mode, switch plugins row by row, and keep several assistants side by side.

GLM Mode agent preset for zhipuai/glm-5.3-flash: PTC/Code Mode tool presentation, a reasoning loop guard and a tool spin guard (live-incident driven), and GLM-specific auto-compaction at 75%/22%.

Agent preset plugin for DeepSeek Harness that triages every request, delegates implementation to an expert, and schedules parallel work through a dependency-DAG taskboard and a file-based message bus before an independent review gate. Ships with only 11 core experts out of the box; more expert teams (Agency agents EN/ZH, classic collections) are one-click downloads from the settings page over dual GitHub/CDN channels with sha256 verification.

Session instruction anchor for DeepSeek Harness: the first step of a new session injects a freely composed instruction block as a plugin-sourced message, and it is re-anchored after a summarizing compaction, every N turns, or when metered context pressure crosses a threshold. An /anchor command re-anchors on demand without a model call; a settings page manages the paged preset library, injection order, re-anchor source and character limits.

Alibaba Bailian (DashScope) model catalog preset + auto-adapter for DeepSeek Harness — reasoning effort levels, thinking budgets, context windows, auto-adaptation of existing Bailian routes, a configurable context ceiling

JetBrains-style Git tool window for the sidebar: branch tree with ahead/behind badges, commit graph with text/author/date/path filters, per-file line diff, changes list and commit box, plus branch, tag and stash actions, merge, rebase, cherry-pick, reset and fetch/pull/push.

Strips the harness-injected prefix and suffix system prompts so the agent runs on the preset's native prompt alone, for soft-jailbreak catgirl persona simulation: ships a bundled plugin that idempotently deploys the 猫娘 Yuike preset to ~/.dsh/.agent-presets/yuike/ and never overwrites an edited preset.

Adds a reasoning-level editor with OpenAI and Anthropic protocol presets (llm-pi-ai), input-modality switches and a 1M/128K context and max-output preset to model rows in Settings.

Use ChatGPT and Codex subscriptions in DeepSeek Harness with OAuth, quota runway forecast, safe resets, web search, image generation, and Fast mode.

Compares existing DSH sessions or runs controlled A/B trials between agent presets, showing aligned timelines, paired metric deltas, explicit success checks, and exportable reports.

Adobe Lightroom Classic as native dsh tools through the lightroom-py MCP server: read the catalog, apply develop settings and presets, export JPEG/TIFF/PSD/DNG, and write ratings, labels, keywords and IPTC — 15 tools driving the Lightroom Classic installed on the local machine, no cloud round-trip; the Python package is on PyPI and the Lightroom plugin installs with one command.

Infinite canvas for sessions, with regions bound to workspaces and agent presets, pinned session cards, sticky notes and alignment guides.

Inline Command Code plan usage badge in the DSH composer toolbar beside the model selector: 5h and weekly window usage plus remaining credits, clicking the badge opens an upward detail popover with reset countdowns; reuses the commandcode/report remote from @mars-sea/dsh-commandcode-provider.

Expert market for DeepSeek Harness: WorkBuddy expert/expert-team system ported onto native agent presets, 425-expert catalog installed on demand, capability subtraction, and team delegation tools.

Maintains reusable named subagent profiles for DeepSeek Harness and runs them with per-profile model, reasoning effort, and agent-preset settings.

An approval-gated OME engineering delivery preset with session-isolated lifecycle evidence.

Install and call external Agent CLIs (Codex, Claude Code) from DSH without touching the system install. Each CLI gets its own Provider, model, reasoning effort and permission preset, callable directly or as a native subagent.

Three independent language channels for model output — tool-call descriptions (every tool-call card title, including run_code in PTC-style presets), thinking, and replies — each set to follow the web GUI language, force a specific language, or turn off (default: only descriptions, following the GUI). One global runtime-context prompt directive on the host; the client half reports the GUI language; a Settings plugin card switches the modes.