Plugins
Navegue, filtre e instale plugins do DeepSeek-Harness.
370 plugins encontrados
dsh-media-kit
daphne999/dsh-media-kit
Pull frames and audio out of a local video with ffmpeg and hand back plain file paths, so the agent can look at the frames with its own vision. Five tools: probe, evenly spaced frames, one frame at a timestamp, scene-change frames (the practical version of frame-by-frame), and audio extraction. No models, no uploads, no routing guesswork.
dsh-memelens
whatif146/dsh-memelens
Explains Chinese internet memes to an agent: registers a meme_explain tool that returns the meaning, tone, origin and lifecycle of a sticker, which memes can and cannot be combined, occasions to avoid, and both a meme-version and a plain-version reply. Answers come from a local memory store with spacing-effect decay, falling back to a user-configured OpenAI-compatible vision endpoint (BYOK).
dsh-review-outbox
zjukop3/dsh-review-outbox
Evidence-gated acceptance for research reports: claims, citations, data scope and method limits each need a tick plus written evidence before a deliverable can pass, with same-artifact revision history and a standalone SHA-256 receipt that says when the bytes moved.
dsh-wenjian-team
sarsgy-wenjian/dsh-wenjian-team
Long-form manuscript verification team for DSH: installs 10 skills that run a read-only content/structure check, three-tier revision proposals, a human-selection gate, exact replacements and a final re-check; it never edits the manuscript by itself and requires Agent Teams to spawn teammates.
dsh-vision
young-felix/dsh-vision
Ask a vision-capable model about a local image from any session, including sessions whose own model is text-only. On the default path the plugin calls the host's route in-process with ctx.llm.stream(), so the host's credentials stay inside its provider adapters; only the optional direct/custom OpenAI-compatible routes read one key (from the plugin config or an apiKeyRef into ~/.dsh/.credentials.yaml).
dsh-feiyu-vision
xxgj666/dsh-feiyu-vision
Image understanding for DSH: describe an image and read the text in it through Zhipu GLM-4V (needs a Zhipu API key, free tier), with optional person identification through Volcengine Ark and anime-only reverse image search through SauceNAO or trace.moe (each needs its own key); ffmpeg must be installed for GIF/WebP input. Keys and the ffmpeg path are set on the plugin's own settings page.
dsh-pf-stream
weilv-d/dsh-pf-stream
Local video understanding for agents: extracts keyframes adaptively based on motion and pauses, inspects any timestamp at full resolution, and aligns visuals with dialogue from embedded subtitles or configured ASR. Requires ffmpeg and ffprobe; clips without subtitles also need a transcription engine for dialogue (a ~190 MB local whisper model, a cloud API key, or the official voice-input model).
dsh-computer-use
lqq151510/dsh-computer-use
macOS computer use with eight computer_* tools: accessibility-tree and Vision OCR screen reading, actions by element index or visible label that also reach background windows, and CoreGraphics mouse and keyboard events with a resident low-latency daemon for game-grade input sequences; its Swift helpers compile on first use, so there is no runtime dependency.
galadon
lumos-789/galadon
A five-stage engineering loop for agent-built projects — spec, build, verify, ship, learn — with one artifact and one gate per phase, so an agent always knows the single next action. Role policy, project templates and deployment sit behind explicit ports: swap the implementation without rewriting the loop. Its scaffold generator lays out the internal layers of a project — `init --layout {none,webapp,service,plugin} --stack {python,node,none}` — from nine layers sampled out of three unrelated production projects: entry docs, an OpenSpec plan face, a spec truth layer (INDEX, one file per domain, TESTING, CONSTITUTION), a wiki split into knowledge / progress / runbook, the six `.agents/` slots, server+web, scripts/deploy/config, tests+e2e and a gitignored out/. Non-none layouts also gate the directory contract and compile every source root. Ships a Python CLI (init/next/check/ship/new) plus a local command adapter and an SSH/systemd adapter with release directories, an atomic current/previous switch, a deploy lock, a commit-versioned JSON probe and automatic rollback on a failed probe; bootstrap and teardown provision and remove the unit and the reverse-proxy block idempotently and refuse to delete a root they did not create. Includes a runnable example workspace and a skill that keeps the agent on one phase at a time. Requires Python 3.10+ and PyYAML; remote deploys need SSH access, rsync, systemd and Caddy on the target.
dsh-pet-watch
learner202699/dsh-pet-watch
桌宠陪看:定时截屏,让视觉模型点评你正在看的画面,投喂给 dsh-pet 的气泡。Pet watch-along for DSH: screen sampling + a vision model + dsh-pet's bubble.
dsh-crash-guard
zxzc114514/dsh-crash-guard
Intercepts startup crashes, isolates broken frontend plugins, and recovers GUI cleanly with revision-bound quarantine.
dsh-codex-runner
halftin/dsh-codex-runner
Runs local Codex CLI tasks from DSH on Windows with background execution, progress queries, follow-up instructions, completed chat records, and optional supervision; requires Node 22.23.4+, DSH >=0.2.1-rc.3 <0.3.0, and an installed, authenticated Codex CLI.
dsh-fairy-zzz-for-jediliu
jediliu005/dsh-fairy-zzz-for-jediliu
Skin H.D.D Fairy III (Zenless Zone Zero) para DSH. Saudações por horário: modelo escreve 5 linhas/dia (até 20 caracteres); GPT-SoVITS sintetiza localmente. Áudio: today -> latest -> bundled audio. Leitura: marcador final > heurística de tamanho; respostas longas resumidas em 1 frase. MD5: 1 reprodução/resposta; troca de sessão silenciosa. Fork com package/rotas/armazenamento próprios, resumo limitado por taxa, motor POST-only, caches TTS/saudações limitados e logs client revision/decision trace locais. Fork de AssassinVIX/fairy-zzz-plugin (upstream dsh-fairy-zzz); instale só um, mesmo fullscreen DSH. Saudações dinâmicas/resumos falados exigem chave DeepSeek compatível em ~/.dsh/fairy-tts.json ou DEEPSEEK_API_KEY; sem chave, só toca a saudação incluída.
dsh-auto-convert
yimengsuper2432/dsh-auto-convert
Prepara arquivos antes de read, read_image e as ferramentas vision os abrirem. Reduz imagens acima de 2048 px ou 8 MB (uma nova tentativa em 1280 px no mesmo turno), converte BMP/TIFF para PNG e identifica uploads com rótulo incorreto examinando os bytes. Converte docx/xlsx/pptx, PDF com camada de texto, csv/tsv e html em Markdown sem programas externos.
dsh-custom-vision
lightingha/dsh-custom-vision
Usa seu próprio modelo vision compatível com OpenAI no DeepSeek Harness: descreve imagens antes que o modelo principal veja a solicitação e oferece `image_understand` para ler gráficos, capturas, UI e OCR.
dsh-model-manager
jyshangguan/dsh-model-manager
Gerenciador de modelos por função para DeepSeek Harness: atribui modelos aos papéis planning/execution/vision/main, roteia subagent por função e escala para planning sessões que chamam uma skill de planejamento (grilling / grill-me).
dsh-relay-isolation
alannerfranklin/dsh-relay-isolation
Prepara um `DSH_HOME` separado para o relay AgentRouter (perfil mínimo e launcher) e oferece egress de rede controlado ao agente em sandbox.
dsh-auto-vision
thanderming/dsh-auto-vision
Quando o modelo selecionado não consegue ver imagens, usa um modelo de visão do mesmo provider para lê-las e passa a descrição ao modelo de texto.
dsh-video-eyes
fj7937290-sys/dsh-video-eyes
Lê vídeos com um modelo somente de texto: com ffmpeg instalado, amostra quadros em intervalos de 0,1 a 10 segundos e cria folhas de contato com o timecode real de cada quadro; extrai o áudio e o envia ao iTingnao para transcrição com timestamps (requer ITINGNAO_API_KEY) e descreve cada folha com um modelo do host capaz de analisar imagens. Funciona sem outro plugin de visão. Um slider do composer define o intervalo; arquivos somente de áudio são transcritos sem amostragem de quadros.
dsh-local-library
fucaplos/dsh-local-library
Biblioteca local de documentos para DSH Desktop (somente Windows; instalada pelo script PowerShell incluído). Envie os arquivos uma vez e qualquer conversa poderá pesquisá-los e lê-los com quatro ferramentas: library_list / library_search / library_read / library_save. Upload, download e download em lote ficam no painel Settings. Os arquivos permanecem no computador, a menos que você habilite o modelo de visão opcional, que envia imagens ao endpoint configurado.
dsh-better-answer-view
solstice621/dsh-better-answer-view
Insere gráficos, tabelas comparativas e blocos interativos entre parágrafos da resposta, com estado salvo localmente e histórico de revisões.
dsh-youtube-study-notes
jagathsrujan/dsh-youtube-study-notes
Transforma um link do YouTube em notas: busca legendas publicadas, pode ler conteúdo visual com um modelo de visão e pede suas notas para combinar tudo em um documento. Requer python3, ffmpeg e yt-dlp no PATH; se faltarem, o skill para com instruções de instalação.
dsh-prompt-inject
379385796-sys/dsh-prompt-inject
Adaptação MIT do gerenciador de modelos de prompt H1a3x com ConfigForms nativos, salvamento verificado por revisão, seleção global/workspace e compatibilidade confirmada com Windows Desktop RC3/RC5. Não pode ser instalado junto ao original H1a3x/dsh-prompt-inject.
dsh-glm-web-search
czpghost/dsh-glm-web-search
Pesquisa web, leitor web e visão Z.AI (GLM) para DeepSeek Harness, autenticados por login OAuth com e-mail Z.AI em vez de API key.