Security audit
Security checks for vulnerabilities and agentic risk
This is a low-risk documentation workflow skill, though its activation wording is broad and may cause accidental use in unrelated tasks.
Install only if you want a general ontology-style workflow helper, and be aware it may trigger on broad productivity or bug-fix phrasing. It does not appear to run code, install dependencies, access credentials, persist changes, or fetch remote payloads.
The skill description is broad enough to match many ordinary productivity, workflow, and implementation-support requests, which can cause the agent to invoke this skill outside its intended niche. Over-broad routing increases the chance of prompt-space takeover by a mediocre or misaligned skill, reducing policy adherence and causing incorrect or unsafe task handling.
The trigger keywords and sample invocations use highly generic terms such as 'knowledge', 'graph', 'creating', and 'bug fix' without contextual constraints, making accidental or excessive activation likely. In a multi-skill environment, this can hijack unrelated user requests, crowd out more appropriate skills, and create unsafe execution paths if downstream behavior assumes the skill was correctly selected.
The trigger sentence is broad and phrased like a normal help request, which can cause the skill to activate when a user did not intend to invoke it. That increases the chance of misrouting prompts, unexpected workflow takeover, and confusion in multi-skill environments where activation boundaries should be explicit.
This trigger pattern is vague and lacks clear scope, so common user requests for a 'practical workflow' could match even when unrelated to this skill's intended ontology-typed use case. In agent routing systems, that can lead to unintended activation, incorrect task handling, and degraded safety if the wrong skill receives sensitive or high-impact requests.
The README describes activation in a way that lacks boundaries between intended and unintended use, making dispatch ambiguous. This increases the risk of accidental invocation by orchestrators or users, especially because the skill targets common concepts like productivity, workflow, and bug fixing that frequently appear in benign conversations.
The trigger sentence is broad enough to match routine user phrasing, which can cause the skill to activate outside its intended scope. In an agent ecosystem, overbroad activation can route unrelated requests into this workflow, leading to prompt hijacking opportunities, incorrect tool use, or unintended processing of user data under the wrong skill context.
The trigger examples are broad, natural-language phrases that can overlap with ordinary user requests, increasing the chance the skill activates unintentionally. In an agent ecosystem, overbroad invocation can route unrelated tasks into this skill, causing misexecution, confusing outputs, or unintended use of workflow logic in contexts the user did not request.
文件整体以中文撰写并命名为 SKILL.zh-CN.md，但内容中未说明这是用户可选语言版本，也未提供语言选择、切换说明或仅适用于中文用户的明确约束。根据规则，强制特定语言而无用户 opt-in 可能构成语言/locale 策略问题。
技能描述覆盖“work-productivity, ontology, typed, knowledge, graph”以及“practical workflow, artifact, checklist, analysis, or implementation support”等大量宽泛任务类型，缺少清晰边界。这样的描述会让调度器或用户难以判断何时真正需要该技能，增加错误调用、职责重叠和将不适合的任务交给该技能处理的风险。
触发关键词包含非常通用的词，如“typed”、“knowledge”、“graph”、“creating”、“bug fix”，很容易与普通技术讨论重叠，导致技能在不相关场景被误触发。误触发本身通常不是直接安全漏洞，但会扩大技能适用面、增加错误自动化和上下文污染风险，尤其当系统依赖关键词自动路由时更明显。
The skill enables implicit invocation while describing activation in very broad terms (e.g., generic work-productivity, ontology, workflow, checklist, analysis, or implementation support). This increases the chance the agent will auto-select the skill in situations beyond the author's intended scope, which can lead to prompt-scope confusion, inappropriate tool use, or unintended influence over unrelated user requests.
This Chinese-language README contains substantial English instructional content, including the core requirement description, without stating that English output is optional or user-selectable. That can create a language/locale policy issue because the file appears targeted at zh-CN users but does not provide opt-in or justification for the language mismatch.
