审计历史
create-pr - 7 审计
版本对比
各审计版本间的能力与发现变化，最新在前。
2026年7月5日 15:15
The static findings are false positives caused by Markdown backticks, fenced examples, and expected Git or GitHub CLI instructions. No prompt injection, data exfiltration intent, hidden network behavior, or malicious command pattern was found in SKILL.md.
风险因素
⚙️ 外部命令 (29)
🌐 网络访问 (2)
2026年7月5日 15:15
The static findings are false positives caused by Markdown backticks, fenced examples, and expected Git or GitHub CLI instructions. No prompt injection, data exfiltration intent, hidden network behavior, or malicious command pattern was found in SKILL.md.
风险因素
⚙️ 外部命令 (29)
🌐 网络访问 (2)
2026年6月30日 02:11
Static analysis flagged many command and network patterns in SKILL.md. Review confirms the external commands are expected git and GitHub CLI workflow instructions, while the weak cryptography findings are false positives with no cryptographic code in the cited lines. The skill can create or edit remote pull requests, so it carries medium operational risk and should be used with authenticated GitHub accounts only.
能力复核项 (3)
这些是真实的本地能力，对此技能可能是预期行为，因此需要复核，但不会按已确认恶意行为计分。
已忽略静态误报 (1)
这些静态命中已被语义复核判定为误报，或只命中了 schema 定义里的词；这里保留展示用于透明度，但不影响质量评分。
风险因素
🌐 网络访问 (1)
检测到的模式
2026年1月17日 05:45
This is a pure markdown documentation skill containing only guidance for creating pull requests using git and GitHub CLI. All 64 static findings are false positives: the scanner misidentified YAML frontmatter as cryptographic code, markdown code fences as shell execution, and documentation links as suspicious URLs. No executable code, scripts, or dangerous capabilities are present.
风险因素
⚙️ 外部命令 (2)
📁 文件系统访问 (2)
2026年1月17日 05:45
This is a pure markdown documentation skill containing only guidance for creating pull requests using git and GitHub CLI. All 64 static findings are false positives: the scanner misidentified YAML frontmatter as cryptographic code, markdown code fences as shell execution, and documentation links as suspicious URLs. No executable code, scripts, or dangerous capabilities are present.
风险因素
⚙️ 外部命令 (2)
📁 文件系统访问 (2)
2026年1月7日 01:42
This is a prompt-based skill containing only markdown instructions for creating pull requests. No executable code, scripts, or potentially dangerous capabilities are present. The skill provides safe git/GitHub CLI guidance.
风险因素
📁 文件系统访问 (3)
⚙️ 外部命令 (4)
2026年1月7日 01:42
This is a prompt-based skill containing only markdown instructions for creating pull requests. No executable code, scripts, or potentially dangerous capabilities are present. The skill provides safe git/GitHub CLI guidance.