Security audit
Security checks for vulnerabilities and agentic risk
This skill is a markdown-only template library for designing agent skills, with no hidden execution, persistence, credential use, or data transfer behavior found.
Reasonable to install if you want guidance for creating agent skills. Be aware that its activation phrases are fairly broad, so it may appear in general conversations about skills or templates; review and narrow triggers if you need stricter routing.
The activation keywords are very generic terms such as 'create skill', 'skill template', and 'design', which can plausibly appear in normal conversation and unintentionally trigger the skill. In an agent environment, ambiguous invocation can cause the wrong skill to activate, leading to unintended behavior, context confusion, or execution of actions the user did not explicitly request.
The skill description is broad enough to match common requests about creating, improving, or structuring skills, which can cause unintended activation outside a narrowly scoped use case. In an agent environment, over-broad activation increases the chance this skill is invoked when another more appropriate skill should handle the request, potentially leading to prompt hijacking of workflow or unnecessary file-loading behavior.
The trigger keywords include generic phrases such as 'create skill', 'skill template', and 'build a skill' without disambiguation or negative constraints, making accidental activation likely. This is more dangerous in context because the skill instructs the agent to load additional files from the references directory, so a loose trigger can expand the skill's influence over unrelated interactions.
The trigger phrases and description are broad enough to match many ordinary writing requests such as summaries, reports, and analysis documents. In an agent skill system, this can cause the skill to activate unexpectedly, override more appropriate skills or default behavior, and steer outputs into a rigid generation workflow the user did not intend.
The trigger phrases are broad enough to match many normal conversations such as 'I want to' or 'build a', which can cause the skill to activate outside its intended scope. In an agent system, this can override more appropriate skills, force unnecessary multi-turn questioning, and create prompt-routing confusion that may be abused to derail workflows or capture excess user context.
The trigger phrases are broad enough to match ordinary documentation-related requests without clearly scoping when this skill should activate. In agent systems, overly broad activation can cause unintended automatic behavior, including loading references, steering the interaction into a rigid pipeline, or taking over requests that did not require this skill.
The template states that it activates when users request documentation generation, but it does not define boundaries for context, repository scope, or file types. That ambiguity increases the risk of unintended invocation in unrelated conversations, which can lead to unnecessary processing, misapplied instructions, or disclosure of loaded context in environments where skills run automatically.
The trigger phrases in the template are broad enough to match many ordinary user requests, which can cause the skill to activate outside its intended scope. In an agent system, over-broad activation can route unrelated or sensitive content into a reviewer workflow, increasing the chance of unintended instruction application, context hijacking, or disclosure through unnecessary checklist-driven processing.
The skill template recommends activation based on broad technology terms like FastAPI, Pydantic, REST API, endpoint, and dependency injection, but does not define scope controls, negative matches, or precedence rules. This can cause the skill to trigger in loosely related conversations, injecting unsolicited instructions and increasing the chance of context hijacking, incorrect guidance, or conflict with other skills.
The suggested activation words include generic terms such as routes, schema, async, and endpoint without contextual constraints, making accidental activation likely. In an agent skill system, overbroad keyword triggers can expand the skill's influence beyond its intended domain, causing inappropriate instruction loading and reducing reliability or safety of downstream responses.
The phrase "English release" signals a language constraint in the skill documentation, but there is no accompanying opt-in, alternative language support, or explanation that the skill is intentionally region- or language-specific. Under the policy, forcing or presuming a specific language without user choice can be a natural-language policy concern.
The phrase "Avoids understanding偏差" introduces non-English text into an otherwise English document, creating an implicit language inconsistency without any opt-in or explanation. This can conflict with language or locale expectations because the file does not state that mixed-language output is intended or region-specific.
The YAML metadata says steps: 4, which communicates the intended pipeline structure. However, the body defines Step 1 through Step 5, including a final delivery stage, so the documentation actively contradicts itself about the number of steps.
The natural-language instructions mandate a specific documentation style: "Use Google style." Under the policy, forcing a specific language/locale-style choice without user opt-in can be a violation unless the constraint is explicitly justified.
The checklist item "Could use more Pythonic写法" mixes English and Chinese in a way that may force a particular language or locale expectation in generated review output or skill behavior. The file does not document this locale choice or offer a user language preference.
No suspicious patterns detected.
