Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent content-review helper, but its very broad activation phrases can route unrelated review requests into a workflow that reads private drafts and compliance profiles and saves a report file.
Install only if you want a TCMS/content-marketing pre-publication reviewer. Use it with explicit draft paths and compliance profiles, and expect it to create a separate review report under content/drafts. The publisher should narrow the activation phrases so ordinary requests like "review" do not invoke this skill accidentally.
The trigger list contains generic phrases such as 'review', '预审', and '审核' that are common in ordinary user interactions and can cause the skill to activate outside its intended compliance-review context. This increases the chance of unintended routing, where unrelated content is treated as sensitive pre-publication material, potentially exposing drafts or causing incorrect compliance judgments.
The title explicitly labels the file as a Chinese-only description, and the document provides no indication that users may choose another language or that the locale restriction is required for a region-specific purpose. This can violate language/locale policy when a skill appears to force a language without opt-in.
The trigger list contains very broad terms such as “审核”, “审稿”, “review”, and “预审”, which are common across many unrelated workflows. This can cause the skill to activate outside its intended compliance-review scope, leading to misrouting of user requests, inappropriate handling of sensitive content, or unintended interception of tasks meant for other skills.
The read_when triggers are very broad generic review terms such as 'review', '审核', and '预审', which can cause the skill to activate for many unrelated requests. This increases the chance of unintended invocation, unnecessary access to drafts and compliance profiles, and silent execution of its file-reading and file-writing workflow in contexts where the user did not explicitly request this skill.
The README defines trigger phrases in both Chinese and English, but it does not explain whether the skill auto-activates based on either language or how language choice is handled for users. This can create ambiguity around locale behavior without explicit user opt-in or documented language-selection logic.
The skill's output/save behavior is defined later in the document, but the user-facing description does not prominently disclose that it writes a report file to content/drafts/{original-name}-compliance-review.md. Hidden or non-obvious persistence can surprise users, create unwanted artifacts, and in sensitive environments may expose filenames, review outcomes, or workflow metadata on disk.
