Security audit
Security checks for vulnerabilities and agentic risk
This is a transparent Chinese writing workflow, but it saves profiles, draft context, source notes, and preferences on disk, so users should understand the local persistence before installing.
Install this if you want a Chinese technical-writing workflow that keeps reusable local notes. Before use, decide whether saving a global writing profile and per-article materials is acceptable for your projects, especially if drafts involve private code or internal context. Review any companion wenqu skills separately before enabling image generation, publishing, translation, or library collection flows.
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
Referenced artifact was not completely inspected
- `Skill` 工具（调用 wenqu-image / wenqu-review / wenqu-translate / wenqu-library）→ 没有跨技能调用机制就直接 Read 目标技能的 `SKILL.md` 和 `references/` 文件，照着内联执行
The skill explicitly relies on external commands such as git rev-parse and openssl rand, which gives a writing workflow shell-adjacent capability outside its core purpose. Any permission to execute subprocesses increases attack surface, especially in a skill that also consumes potentially adversarial repository content and file names.
The activation phrases are broad enough to match many ordinary requests, increasing the chance that the skill runs in contexts the user did not intend. Over-triggering is dangerous here because the skill has persistent storage and file-management behaviors, so accidental activation can lead to unwanted reads, writes, or workflow branching.
The trigger table includes vague cues like ‘改’, ‘检查’, and ‘看看有没有问题’, which are common in general conversation and can unintentionally invoke a powerful workflow. In this skill, that matters more because activation may lead to loading stored context, searching project files, and orchestrating other skills.
The skill instructs reading and updating a persistent profile under $HOME, creating cross-project user profiling state that is not strictly necessary to draft a single article. Even though it says not to store secrets, persistent preference storage can still accumulate sensitive behavioral data and expands the skill’s access beyond the immediate task boundary.
The migration logic authorizes moving and deleting files in the project, which is broader than content drafting and can alter repository state unexpectedly. Even if intended for housekeeping, delete-and-migrate behavior is risky when driven by inferred file structure and can cause accidental data loss or overwrite of user-managed files.
写作技能的描述未包含仓库配置管理或版本控制操作，但此处明确要求检测并可能修改 .gitignore。即使文档说明需要先询问用户，这仍然是与内容创作不同类别的能力。
The rename-recovery flow performs global grep-style searches across wenqu-skills/*/SKILL.md and may rename directories after a match, which grants project-wide file discovery and mutation capability beyond narrow article editing. Because the skill treats article content and metadata as inputs, this increases the blast radius if identifiers are spoofed or if matching logic is wrong.
This markdown file contains all user-facing instructions in Chinese and does not provide any opt-in, alternative language, or justification that the skill is region- or locale-specific. Under the policy, forcing a specific language without user choice is a natural-language policy violation.
This markdown file presents all operational instructions in Chinese and does not provide an opt-in, alternative language, or explanation that the skill is intentionally restricted to a Chinese-speaking context. Under the policy for natural-language violations, forcing a specific language without user choice is in scope.
The entire skill guidance is written as mandatory Chinese-language instructions, including required wording for how to record sources and evaluation conclusions. Under the policy, forcing a specific language without user opt-in or a documented justification is a natural-language policy violation.
The entire skill file is written as Chinese-only operational guidance and does not indicate that language selection is optional or user-driven. Under the stated policy, forcing a specific language or locale without opt-in is a natural-language policy violation unless clearly justified as region-specific.
The entire skill file is written in Chinese and provides procedural instructions only in that language, with no indication that users may choose another language or that the skill is intentionally limited to a Chinese-speaking context. Under the stated policy, forcing a specific language without user opt-in is a natural-language policy violation.
The reusable prompt hard-codes '请用中文' for output, which removes user language choice and can override user expectations or accessibility needs. In this writing skill, that creates a policy and usability issue rather than direct code-execution risk, but it can still cause improper handling of multilingual requests or downstream publication mistakes.
The workflow instructs writing user and project data to multiple persistent files as part of normal operation, but it does not require a just-in-time warning or consent at the point those writes occur. Users may answer seemingly conversational questions without realizing their responses will be stored on disk in reusable project artifacts.
The skill directs persistent storage of reader/author profiles and preferences in project and global paths, creating durable user profiling beyond what is strictly necessary to draft a single article. Because this data is reused across sessions and may include behavioral preferences and background details, it increases privacy risk and the blast radius of accidental disclosure or misuse.
The instruction to immediately append user-provided background material into a persistent materials index can capture sensitive context, internal project facts, or personal information without an explicit retention warning. Since this content is indexed for later reuse, even casually shared details may become durable records that outlive the original task.
The file expands the skill's behavior to include fetching and storing external web materials, but this collection capability is not clearly bounded in the skill description. That can lead to unexpected network access, ingestion of third-party content, and persistence of externally sourced material without the user's informed expectation.
No suspicious patterns detected.