Security audit
Security checks for vulnerabilities and agentic risk
This skill is a local OpenClaw skill-audit helper that is disclosed, mostly read-only, and proportionate, with some non-blocking scope and language notes.
Install only if you want a Chinese-language helper for auditing OpenClaw skill metadata and directory structure. Use it on scoped skill folders, review generated reports before applying changes, and avoid pointing it at broad private directories unless you are comfortable with local content being summarized in the output.
This markdown file contains only Chinese-language headings and audit text, which may force a specific language experience on users without opt-in. The policy explicitly flags language or locale constraints unless the skill offers user choice or clearly documents a justified regional scope.
The skill advertises operational capabilities that include reading files, writing files, and invoking a shell command (python3 .../scripts/run.py), but it does not declare any explicit tool scope such as permissions or allowed-tools. This creates a policy/expectation gap: a caller or orchestrator may permit broader actions than intended, and users cannot clearly audit what the skill is authorized to do.
The file’s headings and instructions are entirely written in Chinese, which implicitly constrains the skill’s interaction language. There is no indication that the user can choose a different language or that the locale restriction is required for a region-specific purpose.
This markdown file is entirely written as a Chinese-language example output and labels itself as such, with no indication that users may choose another language or locale. The policy for this audit flags natural-language constraints that force a specific language without user opt-in.
The document is entirely written in Chinese and labels the skill as a Chinese-language规范检查 workflow, with no indication that users may choose another language or locale. Under the policy, forcing a specific language without opt-in is a natural-language policy violation unless the locale constraint is explicitly documented and justified.
This manifest contains user-facing natural-language fields almost entirely in Chinese, including the title, category label, summary, sections, use cases, and examples. Because the file does not document that the skill is region-specific or offer language/locale choice, it may violate a language/locale policy requiring user opt-in or justification.
The template is entirely written in Chinese and instructs users to fill content in that language, with no indication that output language should follow user preference or be configurable. In a reusable skill template, this can cause unwanted language coercion, reduce accessibility, and lead downstream agents to ignore user locale or policy expectations.
The report-generation strings are written entirely in Chinese, such as '结果', '摘要', and '下一步', which forces a specific language on users. The file provides no option, configuration, or opt-in mechanism to select another language or locale.
The directory, CSV, and pattern report functions emit fixed Chinese headings and guidance text, such as '扫描报告', '字段概览', and '发现结果'. Because this behavior is enforced across multiple modes without a language choice, it violates the language/locale policy criterion.
The built-in pattern scanning logic searches arbitrary files and directories for secrets, shell-pipe execution, destructive commands, and private URLs even though the skill is described as a frontmatter/metadata repair tool. In a trusted agent workflow, this hidden inspection capability broadens access to repository contents and can expose sensitive snippets in generated output without a clear user expectation or least-privilege justification.
Error and audit text like 'SKILL.md 缺少前置 frontmatter', '规范检查', and '文件完整性' is presented only in Chinese. There is no indication that the tool is region-specific or that users can opt into this locale.
The script dispatches to multiple generic report-generation modes based on spec-controlled values, extending well beyond the declared purpose of repairing frontmatter, metadata, and skill structure. This scope expansion is dangerous because it enables repurposing the skill as a general auditing and reporting tool, which can cause unintended data inspection and capability creep in environments that trust the skill's narrower manifest.
Natural-language policy review applies to all file types. The README uses Chinese throughout and does not indicate that the skill is region-specific or that users may choose another language, which can amount to an implicit language constraint without opt-in.
This README is a markdown file, so vague-trigger review applies. The example triggers '检查这个 skill 的 frontmatter 合不合规' and especially '修 requires 和 metadata' are short natural-language phrases without explicit activation boundaries, exclusions, or negative examples, which could overlap with ordinary editing requests rather than clearly constrained skill invocation.
The frontmatter description and title/body text prescribe the skill's behavior in a mixed Chinese/English presentation, but do not indicate that users may choose their preferred language. Under the policy, forcing or assuming a language/locale without opt-in can be a natural-language policy violation.
The code can inventory arbitrary directories, extract Markdown headings, and summarize CSV/TSV datasets. These are general audit/reporting functions that are not obviously required to fix OpenClaw skill frontmatter, metadata, requires/install fields, or directory layout.
No suspicious patterns detected.
