Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-focused OpenAPI helper skill with no executable code, persistence, credential handling, or hidden data access.
Safe to install for OpenAPI/Swagger documentation help. Be aware it may activate on broad API-related prompts because implicit invocation is enabled and trigger keywords are loose; invoke it explicitly when you want this workflow.
The trigger phrases are generic and template-like, which increases the chance the skill is invoked when a user did not clearly request OpenAPI documentation help. In an agent environment, unintended invocation can route user data or workflow control into the wrong skill, causing confusion, irrelevant actions, or unsafe downstream behavior if the skill performs substantive automated guidance.
The trigger phrases are generic and can match broad, ordinary user requests rather than a clear request to invoke this specific skill. That raises the chance of unintended activation, which can cause the agent to apply the wrong workflow, produce irrelevant outputs, or expose users to unnecessary processing paths without explicit intent.
The skill description and use conditions are broad enough to activate on many generic API, Swagger, or documentation-related requests, which can cause the wrong skill to be selected in contexts where more specific or safer handling is needed. Overbroad routing is dangerous because it can misdirect user requests, produce irrelevant artifacts, and increase the chance that downstream actions are taken under incorrect assumptions.
The trigger keywords include generic phrases like 'openapi', 'swagger', 'api documentation', and 'rest api' without qualifiers, making accidental or excessive invocation likely. In a skill-routing system, such unconstrained triggers can overshadow other more appropriate skills and lead to incorrect assistance being applied to broad classes of requests.
该技能文件整体以中文编写，文件名也表明为 zh-CN 版本，但文中未说明这是用户可选的本地化版本，也未提供语言选择或 opt-in 说明。按规则，强制特定语言而无用户选择可能构成语言/地区策略问题。
L04 将技能适用条件描述为“当用户提出 software-and-data, openapi, swagger, api documentation, rest api，或需要围绕该需求获得实用流程、产物、检查清单、分析或实现支持时使用”。其中包含多个非常常见的技术话题词和宽泛的“需要…支持”表述，但没有说明必须满足哪些组合条件、上下文限制或不适用情形，容易覆盖普通对话。
The keyword list and example triggers are overly generic and lack guardrails, which can cause the skill to activate in unrelated conversations. In an agentic system, overbroad activation increases the chance that the model applies the wrong workflow, exposes unnecessary capability surface, or follows documentation-oriented instructions in contexts where they are not appropriate.
The default prompt is extremely broad and generic, increasing the chance that the skill is invoked for routine software or API-related requests even when the user did not clearly ask for it. This can cause prompt-scope overreach, inject unrelated instructions into conversations, and create opportunities for accidental privilege expansion or user confusion.
Enabling implicit invocation without tight trigger constraints allows the skill to be auto-selected in a wide range of contexts, including ones only loosely related to API documentation. In practice this can lead to unintended tool routing, overbroad instruction injection, and reduced assurance that the user explicitly intended to use this capability.
The trigger sentence begins with a very broad phrase ('Help me ...'), which can cause the skill to activate on generic user requests that are not actually asking for OpenAPI or Swagger support. Over-broad activation increases the chance of misrouting conversations, causing the agent to apply this skill in unintended contexts and potentially produce irrelevant or unsafe guidance based on weak matches.
The trigger sentence is ambiguous and insufficiently scoped, so it may match vague workflow requests without clearly tying them to API documentation tasks. This ambiguity can lead to unintended invocation, reducing reliability and increasing the risk that the skill handles requests outside its intended boundaries.
