context-degradation 監査レポート.

警告

• Permission surface may require sandboxing
• SKILL.md excerpt appears truncated at 'When th', making it difficult to verify the complete document is well-formed.
• The detection script explicitly relies on simulated U-shaped attention curves and heuristic token/poisoning estimates, which limits production reliability without real model internals or classifiers.
• SKILL.md does not clearly document setup, installation, or invocation details for the provided scripts and reference materials.
• No explicit limitations or safe operating boundaries section is present, though some boundaries are implied by the activation guidance.
• Quality score needs review
• Permission surface needs review: secrets or environment access, filesystem or document access
• Permission surface: secrets or environment access, filesystem or document access

方法

このレポートは公開メタデータ、AI レビュー、リポジトリの鮮度、インストール準備、OpenAgentSkill イベント、品質スコア、信頼チェック、Agent セーフティゲートを統合します。完全なソースコード監査ではありません。