Security audit
Security checks for vulnerabilities and agentic risk
This is an OpenAPI/Swagger documentation helper with somewhat broad trigger wording, but it contains no hidden code, credential access, persistence, or destructive behavior.
Installers should be aware that the skill may activate on broad API-documentation wording because implicit invocation is enabled. It is otherwise a low-risk documentation helper; review generated OpenAPI specs or code changes before applying them, as with any agent-produced development output.
The trigger phrases are broad and partially generic, which can cause the skill to activate for loosely related requests rather than explicit OpenAPI/Swagger documentation tasks. In an agent environment, over-broad routing can lead to unintended handling of user prompts, incorrect tool selection, and prompt-scope confusion that may expose users to irrelevant or lower-quality actions.
The trigger sentences are generic and loosely scoped, which increases the chance the skill will activate for requests that only partially relate to OpenAPI documentation. In an agent environment, overbroad activation can cause unintended tool selection, irrelevant guidance, or interference with higher-priority skills, especially because the phrasing is templated and not semantically precise.
The document is specifically a Chinese-language README ('README.zh-CN.md') and references separate English and Chinese variants, but this file itself does not state that language selection is optional or user-driven. Under the policy, locale-specific behavior should either offer explicit user opt-in or clearly justify the constraint.
The manifest description says to use the skill when a user 'needs a practical workflow, artifact, checklist, analysis, or implementation support for this requirement,' which is a very broad activation condition layered onto several generic terms. While OpenAPI-related keywords are specific, this phrasing makes the boundary of when the skill should activate less clear and could cause unintended invocation for general software-help requests.
The example triggers begin with 'Help me' and 'I need a practical workflow,' both of which are common conversational phrases and not sufficiently distinctive on their own. Even though the rest of each sentence adds domain context, these examples model vague invocation patterns rather than clearly bounded triggers.
The description states multiple broad domains and task types without clearly defining activation boundaries. Ambiguous activation logic makes accidental invocation more likely, which can steer the agent into using this skill when a general-purpose response would be safer and more appropriate.
The trigger keywords are broad and overlap with ordinary technical discussion terms such as 'openapi', 'swagger', and 'rest api'. This can cause the skill to activate in contexts where the user did not intend to invoke a specialized workflow, increasing the chance of irrelevant instruction injection into normal conversations.
The skill enables implicit invocation while exposing a very broad description and default prompt covering generic software, data, OpenAPI, Swagger, API documentation, workflows, artifacts, checklists, and analysis. This increases the chance the agent will auto-select the skill in loosely related conversations, potentially causing unintended prompt injection surface expansion, context leakage into the skill, or incorrect tool routing.
The trigger sentences are broad enough to match ordinary user phrasing such as 'help me' or 'I need a practical workflow', which can cause the skill to activate outside its intended scope. Over-broad activation is dangerous because it can steer conversations unexpectedly, invoke irrelevant workflow behavior, and create opportunities for prompt-routing abuse or accidental disclosure of context to the wrong skill.
The example trigger phrases are generic, repetitive, and do not distinguish positive from negative cases. Without concrete examples of when not to activate, implementers or routing systems may over-match vague requests and invoke the skill unnecessarily.
The invocation examples are presented only in English and do not indicate whether other languages are supported or how users can choose their preferred language. Under the stated policy, forcing a specific language without opt-in can be a natural-language policy concern when the skill's activation depends on those phrases.
No suspicious patterns detected.
