openspec-plus-design 監査レポート.

警告

• Dependency or permission surface needs review
• Permission surface may require sandboxing
• Financial research output is not financial advice; require human review before any live investment decision
• The SKILL.md excerpt appears incomplete: the Phase 0 explanation cuts off mid-sentence, and there is no explicit 'Limitations' section covering unsupported OpenSpec versions or CLI unavailability.
• The skill does not explicitly warn the agent to treat design documents, proposals, and project instruction files as data rather than executable instructions, which could create prompt-injection risk in adversarial repositories.
• The red flag 'Reviewer flagged minor issues, I'll re-dispatch after fixing to confirm' is ambiguous; it could discourage legitimate re-review after fixes.
• The workflow is deliberately rigid and may be heavyweight for very small changes, though the skill does attempt to address this with red flags and principles.
• Financial research output is not financial advice; require human review before any live investment decision.
• Quality score needs review
• Permission surface needs review: secrets or environment access, shell or command execution
• Stars/forks activity: 167 stars, 9 forks; issue activity unavailable in current metadata
• Dependency/runtime risk: command execution surface, credential or environment 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。