Discover verified DSH plugins

@skylarking/dsh-chat-renderer

Chat Renderer

Installable chat content renderer extensions for DeepSeek Harness

dsh-virtual-eng-team

Virtual engineering team: turn an agent into a full engineering team—CEO rethinks the product, engineering manager locks down the architecture, designer catches AI slop, reviewer finds production bugs, QA uses a real browser, security officer runs OWASP+STRIDE, and release engineer opens PRs; 23 exp

dsh-test-desktop-app

Test Desktop App

End-to-end testing for Electron applications: connect to the real Electron renderer via the Chrome DevTools Protocol (CDP)—DOM inspection, JS evaluation, clicking, screenshots, DPR multipliers, genuine right-click event triggering, xterm layout delays, and other pitfalls. Inspired by bholmesdev/hubb

Taste Review

Aesthetic judgment call: for ambiguous UI refinement, copy, naming, or layout decisions, initiate a taste-driven judgment call—list files, weigh several options, provide a recommendation, and list alternatives; match the depth to the decision (design calls may take a few paragraphs, naming calls do

dsh-social-simulation

Social Simulation

Million-agent social simulation: a scalable, open-source social media simulator in which LLM agents realistically emulate up to one million users on Twitter/Reddit, with 23 actions (follow/comment/repost), interest- and popularity-based recommendation algorithms, and a dynamic environment; for study

dsh-skill-optimization

Skill Optimization

Treat skills as trainable parameters: train agent skills like neural networks (epochs/batchsize/learning rate/validation gates, without modifying model weights)—a rollout→reflect→aggregate→select→update→evaluate loop; accept candidate edits only when they strictly improve the held-out validation sco

dsh-self-improving-agent

Self Improving Agent

Self-improving AI agent: built-in learning loop—automatically creates skills from experience, improves with use, cross-session recall through FTS5 session search + LLM summaries, Honcho dialectical user modeling, cross-platform gateway (Telegram/Discord/Slack), scheduled automation, parallel sub-age

dsh-parallel-agent-ade

Parallel Agent Ade

并行 agent 开发环境：Codex/ClaudeCode/OpenCode/Pi 并排跑，各在独立 git worktree——一个 prompt 扇出 5 个 agent 比较合并胜者；移动伴侣监控、终端分屏、Design Mode（点击 UI 元素送 HTML/CSS/截图进 prompt）、GitHub/Linear 原生、SSH worktree、AI diff 标注。受 stablyai/orca（49k★ MIT）启发。

Model Gateway

Model routing gateway: connect to 290+ AI providers through a single endpoint (90+ free, 1,200+ models), fully compatible with Claude Code/Codex/Cursor/Cline/Copilot—with quota-aware automatic fallback, 19 routing strategies, 15–95% token savings through RTK+Caveman stacked compression, aggregated f

dsh-knowledge-brain

Knowledge Brain

Knowledge brain layer: search gives you raw pages, while the brain gives you cited answers — synthesis layer (actual answer + sources + explicit indication of what it does not know) + self-linking knowledge graph (extracts entity edges with zero LLM calls) + gap analysis + 24/7 dream cycle (nighttim

Geo Seo

GEO-first AI 搜索优化：为 AI 搜索引擎（ChatGPT/Claude/Perplexity/Gemini/AI Overviews）优化网站——citability 引用就绪评分、AI crawler 访问分析、llms.txt 生成、品牌提及扫描、Schema 标记、E-E-A-T 内容质量、5 并行子代理全审计。受 zubair-trabzada/geo-seo-claude（9.4k★ MIT）启发。

dsh-desktop-agent-gui

Desktop Agent Gui

Agent desktop orchestration interface: Turn CLI agent orchestration into a visual desktop experience—with zero-configuration setup (wizard), intelligent chat (streaming Markdown/@agent routing/inline skill cards), multi-channel management, cron-based scheduled automation, local skill management, and

dsh-cloud-well-architected

Cloud Well Architected

Cloud Well-Architected six-pillar review framework: reliability (user experience goals/SLOs/redundancy/horizontal scaling/observability/graceful degradation/recovery testing/blameless retrospectives), cost optimization, operational excellence, performance optimization, security, and sustainability—d

dsh-channel-assistant

Channel Assistant

Channel-based personal assistant: a personal AI assistant for a single operator, connecting models/tools/messaging channels/companion apps through one Gateway—WhatsApp/Telegram/Slack/Discord/Signal/iMessage and other channels as the interface, Control UI/CLI/TUI, a pairing security model (unknown se

dsh-autonomous-research

Autonomous Research

Autonomous research agent: Give an AI agent a small, real LLM training setup and let it run experiments overnight—modify code → run a fixed 5-minute training session → check whether val_bpb improved → keep or discard → repeat (about 12 experiments/hour, roughly 100 overnight); single-file modificati

v0.3.5

@sandbaseai/dsh-skills

Skills

Install SandBase Agent Skills into DeepSeek Harness projects.