downy
bensenescu/downy
Downy is a self-hosted multi-agent orchestration framework built on Cloudflare Durable Objects, allowing users to coordinate teams of specialized AI agents from any device. It supports flexible model backends including Workers AI, OpenRouter, and ChatGPT Plus/Pro subscriptions.
196 34MITUpdated 2026-05-06
70
GoodOverview
Downy enables users to build and manage a collaborative team of AI agents, each featuring distinct personalities, designated workspaces, and specific toolsets. Designed to be deployed effortlessly on Cloudflare via Alchemy, the system isolates each agent and subagent as an independent Durable Object. Users can interact with their agent crew securely from anywhere behind Cloudflare Access, utilizing cost-effective default models or proxying existing subscription credentials without incurring traditional per-token API costs.
Capabilities
- ▸Multi-agent team management with individual workspaces and personalities
- ▸Cloudflare Workers and Durable Objects integration
- ▸Model integration supporting Workers AI, OpenRouter, and ChatGPT subscriptions
- ▸Secure remote access via Cloudflare Access authentication
- ▸Web-based dashboard for managing agents, tools, and background tasks
Best for
Building and managing a custom multi-agent team for automated workflows and background tasks, Running personal AI agents hosted securely in a private Cloudflare infrastructure, Leveraging ChatGPT Plus/Pro subscriptions or OpenRouter models through a self-hosted web interface
Works with
CursorClaude Code
Evoa Score breakdown= Σ (score × weight)
70Task usefulness20%88 → +17.6
実タスクにどれだけ役立つか（機能の豊富さ・用途の明確さ）。 — AIによるcapabilities/use-cases解析
Code quality15%85 → +12.8
実装・指示の品質。 — AIによるSKILL.md/README解析
Maintenance15%70 → +10.5
リポジトリがどれだけ活発に保守されているか。 — GitHub 最終push日時の新しさ
Documentation12%42 → +5.0
ドキュメントの充実度・分かりやすさ。 — README/独自要約の情報量
Security15%50 → +7.5
危険・不審な挙動が無いか。 — AIによるセキュリティレビュー
Originality10%90 → +9.0
ありふれたラッパーではない独自性。 — AIによる独自性判定
Popularity8%61 → +4.9
コミュニティの採用度。 — GitHub Stars/Forks（対数スケール）
Compatibility5%50 → +2.5
対応AIエージェントの広さ。 — AIによる対応エージェント判定
Weighted total69.8 / 100
ライセンス不明/制限あり（Red）のSkillは総合スコアに0.85倍の補正を適用します。 ランキングはこのScoreのみで決まり、広告で変わりません。 算出方法の詳細 →
Security considerations
The application relies on Cloudflare Access for network-level authentication. However, because agents possess self-editing features and tool access, the system is inherently prone to prompt injection risks similar to other agentic frameworks, and users should carefully manage permissions.
Categories
Summary and analysis are original content generated by AI Skills Rank. The skill's source text is not reproduced here — view it on the linked repository.