context-optimization 审计报告.

警告

• Permission surface may require sandboxing
• The SKILL.md excerpt is incomplete at the end, so the full masking rules and edge cases are not fully visible in the provided snippet.
• The Python utilities use heuristic token estimation and simple summarization; the code itself notes that production systems should use model-specific tokenizers and LLM-based summarization.
• The skill does not explicitly address handling untrusted or adversarial content inside tool outputs during masking or compaction, though the stated rules mitigate some risk.
• 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

本报告综合公开元数据、AI 审查输出、仓库活跃度、安装就绪度、OpenAgentSkill 事件、质量评分、信任检查和 Agent 安全门槛；它不是完整的源代码安全审计。