Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with broad routing language, but it does not include hidden code, credential handling, persistence, or destructive behavior.
Install this only if you want a general Agent Browser workflow helper. Be aware that its trigger words are broad and implicit invocation is enabled, so it may appear for ordinary browser, automation, CLI, or bug-fix requests; publishers should narrow those triggers for cleaner routing.
The skill description and usage scope are broad enough to match many ordinary requests unrelated to a narrowly defined browser workflow helper. In an agent routing system, this can cause unintended invocation, increasing the chance that the skill handles prompts outside its validated boundaries and produces unsafe, irrelevant, or over-privileged automation guidance.
The trigger keywords include highly generic terms such as 'browser', 'automation', 'cli', and 'bug fix', which are common across many unrelated user requests. This creates a significant prompt-routing collision risk, allowing the skill to be selected in contexts it was not designed for, potentially overriding more appropriate safeguards or task-specific skills.
The trigger sentence is written as a very broad, everyday help request rather than a narrowly scoped invocation cue. In an agent ecosystem, this can cause accidental activation outside the intended context, leading the skill to engage on unrelated tasks and potentially expose users to unintended automation or misleading workflow guidance.
The trigger phrases are generic and overlap with common user requests about browsers, automation, workflows, and productivity, which raises the chance the skill is invoked when the user did not explicitly intend it. In an agent setting, unintended invocation can route tasks into the wrong workflow, causing inappropriate actions, misleading outputs, or accidental use of browser/headless automation capabilities.
The trigger phrases are broad, generic, and closely resemble normal user requests such as asking for help or a practical workflow. This can cause the skill to activate unintentionally in unrelated contexts, which is risky for a browser/headless automation helper because it may steer conversations toward automation behaviors the user did not explicitly request.
The example trigger sentences are malformed and truncated, so they do not establish clear activation boundaries or reliable examples of intended use. Ambiguous examples increase the likelihood of accidental matches, user confusion, and inconsistent routing behavior in systems that rely on pattern-like trigger text.
The description states broad usage conditions such as work-productivity, browser, headless, and automation without defining firm activation boundaries. Ambiguous invocation logic can cause the skill to be selected outside its intended scope, producing incorrect assistance or interfering with safer, more appropriate skills.
The trigger keywords are overly broad, including generic terms like browser, automation, and cli, which can cause the skill to activate for unrelated requests. This increases the chance of unintended routing or privilege expansion, especially in agent ecosystems where skill selection influences downstream actions and user trust.
The default prompt trigger phrase is extremely broad and overlaps with common user language such as 'help me' and generic productivity/browser requests. With implicit invocation enabled, this can cause the skill to activate in unintended contexts, creating prompt-routing confusion and increasing the chance of unauthorized or surprising execution of the skill’s instructions.
The invocation guidance includes generic request forms like asking for 'a practical workflow' and a direct-use pattern that can match many normal user requests. This increases the risk of overbroad routing or unintended skill selection, which is particularly relevant for a browser/headless automation helper because accidental activation could steer users into unnecessary automation steps or unsafe operational suggestions.
The file inventory explicitly labels separate English and Chinese instructions and guides, but the README does not state how the user selects a language or that language choice is optional. This can imply a fixed locale path rather than an explicit user opt-in, which may conflict with language/locale policy expectations.
This Chinese README contains a full English sentence block in the main description, which can impose an unexpected language experience on users of the zh-CN version. The file does not indicate that language choice is optional or user-selectable.
该技能文件整体以 zh-CN 语言版本呈现，内容中未说明用户可选择其他语言或根据用户偏好切换。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，属于自然语言策略风险。
No suspicious patterns detected.
