Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-style helper skill with broad routing language, but it does not contain code, hidden execution, credential handling, persistence, or data exfiltration behavior.
Installers should be aware that this skill may activate too broadly for general browser, automation, CLI, or bug-fix requests. It is otherwise a low-risk documentation skill; review the prompts if precise skill routing matters in your environment.
The manifest description uses expansive activation language such as 'Use when a user asks for work-productivity, agent-browser, browser, headless, automation,' which is too general for a safe skill boundary. Overbroad scope increases unintended invocation risk and may cause this skill to supersede more appropriate or safer skills, especially in common browser or automation requests.
The trigger list is very broad and includes generic terms like 'browser', 'automation', 'cli', and 'bug fix', which are likely to match many unrelated requests. This can cause the skill to activate outside its intended scope, leading to confusing routing, accidental takeover of user tasks, or unsafe application of browser-automation guidance in contexts where it was not requested.
The default_prompt uses a very broad invocation phrase and includes generic terms like 'help me' and broad workflow categories, which can cause the skill to activate in contexts unrelated to the user's actual intent. Because implicit invocation is enabled, this increases the chance of unintended routing, prompt injection surface expansion, or accidental application of this skill's behavior to unrelated conversations.
The trigger sentence is extremely broad and malformed, making it likely to activate on generic user requests rather than clear skill-specific intent. In an agent environment, over-broad activation can route unrelated prompts into this skill, causing unintended browser/automation guidance to be applied in the wrong context and increasing the chance of unsafe or confusing execution paths.
The README defines trigger sentences using very broad natural-language phrases tied to common terms like 'help me', 'practical workflow', and 'agent-browser', without meaningful scope restrictions. In an agent-routing environment, this can cause unintended activation for loosely related requests, leading the agent to apply browser-automation guidance in contexts the user did not intend, which increases the chance of unsafe actions, overreach, or confusion.
The trigger phrases are broad and generic enough to match many routine requests about browsers, automation, CLI, or productivity, which can cause unintended skill activation. In an agentic workflow, overbroad activation increases the chance that this skill handles requests outside its intended scope, leading to incorrect automation behavior, misleading outputs, or unsafe browser-task assistance in contexts the user did not explicitly intend.
The usage condition in the skill description is overly broad and mixes several generic categories like work-productivity, browser, headless, and automation without clear boundaries. In a routing system, this ambiguity can make the skill eligible for many unrelated tasks, increasing the chance of misfires and potentially exposing users to irrelevant or risky workflow suggestions.
The trigger keywords include very broad terms such as "browser", "automation", and "cli", which are common in many unrelated conversations. This can cause accidental invocation of the skill outside its intended scope, leading to inappropriate routing, confusing outputs, or unintended use of browser-oriented workflows in contexts where they do not fit.
This trigger is ambiguous because it does not clearly state when the skill should be selected versus other general productivity or automation skills. Ambiguous routing increases accidental invocation, which can waste agent steps, produce irrelevant outputs, and weaken safety controls that depend on accurate skill selection.
Despite being a zh-CN README, several core user-facing sections and trigger phrases are written in English, including the title, demand description, workflow description, and invocation examples. This can amount to an implicit language constraint without user opt-in, which conflicts with language/locale policy expectations.
No suspicious patterns detected.
