Security audit
Security checks for vulnerabilities and agentic risk
This skill is a documentation workflow helper for OpenAPI/Swagger tasks and does not include code execution, credential access, persistence, or hidden data movement.
Installers should expect a lightweight OpenAPI/Swagger documentation assistant. The main caveat is routing: because its activation text is broad, users should invoke it explicitly for REST API documentation work and avoid relying on it for unrelated software or developer-experience questions.
The trigger sentence is so broad and malformed that it can match a wide range of ordinary user requests, causing the skill to activate outside its intended scope. Over-broad activation increases the chance of prompt hijacking at routing time, misapplication of specialized instructions, and accidental disclosure or transformation of user input in unintended contexts.
The trigger phrases are generic and structurally broad enough that normal user requests about OpenAPI, Swagger, or practical workflows could invoke the skill unintentionally. In an agent system, overbroad routing increases the attack surface for prompt injection, mis-execution, or context confusion because the skill may activate outside the user’s actual intent boundary.
The documented trigger phrases are broad, templated, and partially malformed, which can cause the skill to activate in situations beyond the author's likely intent. In an agentic environment, ambiguous activation increases the chance of accidental routing to this skill, causing inappropriate handling of user requests or creating opportunities for prompt-scope confusion with other skills.
The manifest description uses broad activation language such as 'Use when a user asks for software-and-data, openapi, swagger, api documentation, rest api,' which can cause the skill to be selected for loosely related requests. Overbroad routing is a real security and reliability issue because it increases the chance this skill handles requests outside its intended scope, potentially crowding out more appropriate skills or causing unintended data handling.
The trigger keywords include broad terms like 'software-and-data' and 'developer experience' without scope constraints, making accidental or inappropriate activation more likely. In agent ecosystems, imprecise triggers can become a security boundary issue because they steer user requests into tools or workflows that were not intended for the actual task.
The skill description and usage conditions are broad enough that ordinary requests about software, APIs, or documentation could trigger the skill even when the user did not specifically want OpenAPI-generation help. Overbroad triggering can cause inappropriate routing, irrelevant guidance, or unintended exposure of specialized instructions in contexts where they do not fit.
The trigger section relies on generic keywords like 'software-and-data', 'openapi', 'swagger', 'api documentation', and 'developer experience' without defining boundaries or counterexamples. This makes accidental invocation more likely, which can degrade system behavior and cause the wrong skill to handle common requests.
The skill enables implicit invocation while providing only very broad activation criteria around OpenAPI, Swagger, API documentation, and related workflow help. This can cause the agent to trigger the skill in loosely related contexts without clear user intent, increasing the chance of unwanted prompt injection exposure, inappropriate tool use, or accidental influence over unrelated conversations.
The invocation phrase is ambiguous because it does not clearly delimit when the skill should be used versus when a general assistant should respond. Ambiguous routing can cause unintended execution of the skill on loosely related requests, which weakens trust boundaries and may let adversarial users steer the system into an inappropriate capability path.
The README declares both English and Chinese user-facing guides, but it does not state how language selection occurs or that the user can choose their preferred language. This can create ambiguity around locale behavior and may conflict with policies requiring language choice rather than implicit locale assumptions.
In this Chinese README, the documented trigger phrases are only provided in English, which can impose an implicit language requirement on users. The file does not indicate that users may invoke the skill in Chinese or choose their preferred language.
The example triggers are ambiguous and malformed, which reduces activation specificity and makes routing behavior less predictable. While this is less severe than broad keywords, unclear examples can still contribute to accidental invocation and make it harder for downstream systems or reviewers to understand the intended boundary of the skill.
Several evidence entries and linked titles are in Chinese, but the file does not explain any language or locale expectations for users. This may create an implicit language constraint without offering the user a language choice or documenting why that locale is required.
