Security audit
Security checks for vulnerabilities and agentic risk
This is a simple unit-test assistance skill with overly broad auto-invocation keywords but no hidden code, credential access, persistence, or destructive behavior.
Before installing, consider narrowing or disabling implicit invocation because common words may trigger the skill when you did not intend it. From a security standpoint, the artifact appears proportionate for unit-test support and does not add privileged runtime behavior.
The explicit keyword list contains highly generic single-word triggers like 'add', 'unit', 'tests', 'level', and 'type', which are likely to appear in many normal conversations. In an agent routing system, this materially increases the chance of false activation, causing misrouting, context confusion, and possibly execution of this skill in situations unrelated to unit testing.
The trigger terms in the metadata are extremely broad (add, unit, tests, labels) and are not scoped to a specific intent or phrasing. This can cause the skill to activate during unrelated conversations, which may steer the agent into irrelevant behavior, override a more appropriate skill, or create unintended disclosure of workflow content.
The trigger sentence uses very generic terms around 'add', 'unit', and 'tests', which can cause the skill to activate for unrelated requests that merely mention those common words. In an agentic system, unintended activation can route user requests into the wrong workflow, causing confusing guidance or inappropriate actions, even though this skill appears informational rather than overtly dangerous.
The trigger phrases are broad enough to match ordinary help requests such as asking for practical workflow support or generic unit-test help, which can cause the skill to activate outside a clearly intentional invocation. In an agent system, over-broad activation can misroute user requests, cause unintended behavior, and increase the chance that a user is steered into this skill when they did not explicitly choose it.
The file descriptions explicitly label the main skill specification as English and this README as Chinese, but there is no statement that users can choose their preferred language or that the locale split is required for a region-specific purpose. This can conflict with language/locale policy expectations when a skill appears to prescribe language variants without opt-in.
The skill description includes broad activation language such as 'general-help', 'add', 'unit', and 'tests', which are common terms in many unrelated requests. This can cause unintended invocation of the skill, leading the agent to apply the wrong workflow or expose irrelevant guidance in contexts where the user did not actually request unit-test support.
该技能文件整体以中文呈现，文件名也表明为 zh-CN 版本，但内容未说明这是可选语言版本，也未告知用户可以选择其他语言。按规则，若技能强制特定语言而没有用户选择或明确合理的区域限定，属于自然语言策略问题。
The keyword list under the trigger section lacks boundaries, priority rules, or examples of non-matching cases, so common words can match outside the intended workflow. In an agentic environment, ambiguous activation logic increases the risk of accidental invocation and incorrect task routing, which can degrade safety and reliability.
The default prompt and description use very broad phrasing such as 'help me Add Unit Tests' and the metadata says to use the skill for 'general-help' and implementation support, which creates a loose trigger surface. In combination with implicit invocation, this can cause the skill to activate for routine development conversations where the user did not clearly request this helper, increasing prompt-scope confusion and unintended delegation.
Enabling allow_implicit_invocation without additional context constraints allows the skill to be auto-invoked during loosely related conversations. Because this skill is framed broadly around practical workflow, checklist, analysis, and implementation support, unintended activation could steer agent behavior, expose unrelated context to the skill, or cause actions to be taken under the wrong capability.
The skill's activation signals include very generic keywords such as 'general-help', 'add', 'unit', and 'tests', which can match many unrelated user requests and cause the skill to trigger outside its intended scope. Over-broad invocation increases the chance of inappropriate routing, prompt interference with other skills, and accidental exposure of users to instructions that are irrelevant or lower quality for the task at hand.
The trigger sentence 'Help me Add Unit Tests' is broad natural language that closely resembles ordinary user phrasing, so it may activate in many routine conversations where the user is not actually asking to invoke this specific skill. This ambiguity can lead to unintended skill selection and reduced reliability of task routing, though the content of this skill is operationally low risk.
No suspicious patterns detected.
