Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with broad activation wording, but it does not install code, access data, persist, or execute browser automation itself.
Installers should be aware that this skill may activate for broad browser or automation requests. Prefer invoking it explicitly by name when you want workflow guidance, and review its outputs before applying any generated browser-automation changes.
The trigger sentence begins with a very broad everyday phrase ('Help me') and then appends a long requirement description, which can cause accidental invocation from unrelated user requests. In a browser/automation-oriented skill, over-broad activation is more dangerous because it may steer general conversations into workflow or automation behavior the user did not explicitly request.
The trigger sentence is so broad and malformed that it could cause the skill to activate in loosely related conversations rather than only when the user explicitly wants this workflow. Over-broad activation increases the chance of unintended instruction injection into unrelated tasks, which is especially risky for a browser/headless automation helper that may influence higher-impact actions.
This trigger remains ambiguous because it describes a generic need for a 'practical workflow' using broad ecosystem terms instead of a constrained use case. In context, the skill targets browser and automation workflows, so accidental activation could steer the agent into unnecessary or unsafe automation-oriented guidance when the user's intent is only tangentially related.
The README lists invocation phrases under '触发句' that include broad natural-language prompts like 'Help me ...' and 'I need a practical workflow ...'. These overlap with common everyday speech and do not clearly define when this skill should activate versus when it should not, increasing the risk of unintended invocation.
The manifest description says to use the skill when a user asks for broad categories such as 'browser', 'automation', or a 'practical workflow', which are everyday phrases shared by many unrelated tasks. This ambiguous activation condition can cause accidental invocation, inappropriate delegation, and policy drift because the skill may handle requests beyond the validated requirement it was designed for.
The trigger list is broad and includes common terms like 'browser', 'automation', 'cli', and 'bug fix', which can cause the skill to activate for many ordinary requests outside its intended scope. Over-broad activation increases the chance of misrouting user tasks to this skill, leading to unsafe or irrelevant automation guidance and reducing the reliability of skill selection.
The skill description says it should be used for broad categories like work-productivity, browser, headless, and automation, without clearly defining exclusions or the precise task boundary. In a shared agent environment, this ambiguity can cause accidental selection of the skill for loosely related requests, increasing the chance of irrelevant actions, poor delegation, or unintended overlap with other skills.
The trigger list includes very broad generic keywords such as "browser", "automation", "cli", and "bug fix", which can match many unrelated user requests. This creates an overbroad auto-invocation surface where the skill may activate outside its intended scope, potentially causing wrong guidance, workflow confusion, or unsafe task routing in an agent system.
The default prompt is overly broad and generic, so it can plausibly match ordinary user requests that mention browsers, automation, productivity, or workflows even when the user did not intend to invoke this skill. That increases the chance of unintended skill activation, causing the agent to apply this skill's instructions or workflow in unrelated contexts and potentially expanding the skill's influence beyond user intent.
Enabling implicit invocation without precise activation criteria allows the platform to auto-select this skill based on loose semantic similarity rather than clear user intent. In a browser/automation-oriented skill, unintended activation is more dangerous because it may steer the agent toward actions, recommendations, or artifacts affecting reliability, setup, or safety-sensitive workflows.
The activation examples are ambiguous and largely restate requirement text instead of defining clear conditions for when the skill should be used. That ambiguity increases the chance of misrouting ordinary productivity or help-seeking prompts into this skill, which is especially risky given the skill covers browser, headless, and automation workflows that may influence sensitive operational actions.
This file is labeled as a Chinese README, but core user-facing content such as the requirement description is presented in English. That can impose a language choice on users without opt-in and may conflict with language/locale expectations for a localized skill document.
No suspicious patterns detected.
