20x
peakflo/20x
Peakflo 20x is a production multi-agent system designed to run parallel AI coding agents with centralized polling, secret management, and human-in-the-loop approvals.
96 16MITUpdated 2026-08-13
77
GreatOverview
This multi-agent architecture coordinates multiple AI coding agents across different codebases using a centralized polling coordinator and unified adapter interfaces. It supports OpenCode SDK, Claude Code, and Codex backends, manages SQLite-backed agent persistence, handles encrypted environment variables via secret brokers, and provides real-time streaming transcripts with human-in-the-loop workflows.
Capabilities
- ▸Multi-agent parallel execution
- ▸Adapter-based backend abstraction for OpenCode, Claude Code, and Codex
- ▸Centralized polling coordinator with deduping and stuck-session detection
- ▸SQLite-backed persistent agent, task, skill, and secret storage
- ▸Encrypted secret management and environment variable injection
- ▸IPC communication channels for agent CRUD and session control
- ▸Human-in-the-loop approval workflows and streaming transcripts
Best for
Running parallel AI coding agents on distinct tasks and codebases, Managing multiple AI agent backends like Claude Code, OpenCode, and Codex under one interface, Enforcing human-in-the-loop approvals for autonomous coding operations, Securely injecting environment variables and secrets into agent execution contexts
Works with
Claude CodeCodexOpenCode
Evoa Score breakdown= Σ (score × weight)
77Task usefulness20%100 → +20.0
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%90 → +13.5
実装・指示の品質。 — AIによるSKILL.md/README解析
Maintenance15%100 → +15.0
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%36 → +4.3
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%85 → +8.5
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%53 → +4.2
コミュニティの採用度。 — GitHub Stars/Forks（対数スケール）
Compatibility5%75 → +3.8
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 total76.8 / 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
Secrets are stored as encrypted blobs in SQLite and injected into execution contexts via a secure broker. Permission modes ('ask' vs 'allow') regulate autonomous tool execution.
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.