技能 shirushi 审计历史
📦

审计历史

shirushi - 6 审计

审计版本 6

最新 中风险

Jun 28, 2026, 08:21 PM

Static analysis flagged many backtick examples as Ruby or shell execution, but the file is Markdown documentation for a CLI workflow. No malicious intent, network access, credential handling, or prompt injection was found. Risk remains medium because the skill directs agents to run external commands that can inspect and modify repository files.

1
已扫描文件
99
分析行数
5
发现项
codex
审计者
中风险问题 (1)
External CLI Commands Can Modify Repository Files
The skill instructs agents to run shirushi commands. Validation and display commands appear safe, but assign and rehash workflows can change document metadata or content hashes in the working repository.
低风险问题 (3)
Static Backtick Execution Findings Are Markdown Examples
The static analyzer flagged inline backticks and fenced examples as Ruby or shell execution. In context, these are documentation examples for the shirushi CLI, not executable code embedded in the skill.
Weak Cryptography Finding Not Confirmed
The scanner reported a weak cryptographic algorithm at the description line, but that line only describes document ID management. The checksum configuration later names mod26AZ, which is an identifier check digit, not a security control.
System Reconnaissance Findings Are Domain-Specific Lookups
The flagged show, scan, and reference examples operate on document metadata within the repository. No evidence found of host enumeration, credential discovery, process listing, or other system reconnaissance.

检测到的模式

Agent-Run External Commands

审计版本 5

安全

Jan 16, 2026, 08:49 PM

This skill contains ONLY documentation (SKILL.md). The static scanner produced 37 false positives by misidentifying markdown code formatting and configuration examples as security threats. No executable code, network calls, or malicious patterns exist. All findings are dismissed as FALSE_POSITIVE.

2
已扫描文件
176
分析行数
1
发现项
claude
审计者
未发现安全问题

审计版本 4

安全

Jan 16, 2026, 08:49 PM

This skill contains ONLY documentation (SKILL.md). The static scanner produced 37 false positives by misidentifying markdown code formatting and configuration examples as security threats. No executable code, network calls, or malicious patterns exist. All findings are dismissed as FALSE_POSITIVE.

2
已扫描文件
176
分析行数
1
发现项
claude
审计者
未发现安全问题

审计版本 3

安全

Jan 10, 2026, 11:47 AM

This is a prompt-based documentation skill containing only a SKILL.md file. No executable code, network calls, or filesystem access patterns are present. The described shirushi CLI tool is a legitimate document ID management system with no malicious indicators.

1
已扫描文件
99
分析行数
0
发现项
claude
审计者
未发现安全问题

审计版本 2

安全

Jan 10, 2026, 11:47 AM

This is a prompt-based documentation skill containing only a SKILL.md file. No executable code, network calls, or filesystem access patterns are present. The described shirushi CLI tool is a legitimate document ID management system with no malicious indicators.

1
已扫描文件
99
分析行数
0
发现项
claude
审计者
未发现安全问题

审计版本 1

安全

Jan 10, 2026, 11:47 AM

This is a prompt-based documentation skill containing only a SKILL.md file. No executable code, network calls, or filesystem access patterns are present. The described shirushi CLI tool is a legitimate document ID management system with no malicious indicators.

1
已扫描文件
99
分析行数
0
发现项
claude
审计者
未发现安全问题