監査履歴
route - 3 監査
監査バージョン 3
最新 安全Jan 16, 2026, 01:07 PM
Pure prompt-based skill with no code execution, file access, network calls, or external commands. The skill defines argument parsing logic and subagent routing behavior for the Claude Router system. All static findings are false positives: metadata URLs flagged as network access, markdown code fences flagged as backtick execution, and routing terminology flagged as C2/crypto patterns.
リスク要因
🌐 ネットワークアクセス (1)
📁 ファイルシステムへのアクセス (1)
⚙️ 外部コマンド (22)
監査バージョン 2
安全Jan 16, 2026, 01:07 PM
Pure prompt-based skill with no code execution, file access, network calls, or external commands. The skill defines argument parsing logic and subagent routing behavior for the Claude Router system. All static findings are false positives: metadata URLs flagged as network access, markdown code fences flagged as backtick execution, and routing terminology flagged as C2/crypto patterns.
リスク要因
🌐 ネットワークアクセス (1)
📁 ファイルシステムへのアクセス (1)
⚙️ 外部コマンド (22)
監査バージョン 1
安全Jan 10, 2026, 09:00 AM
Pure prompt-based skill with no code execution, file access, network calls, or external commands. The skill only defines argument parsing logic and subagent routing behavior for the Claude Router system.