Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-style workflow helper with overly broad activation wording, but it does not ship executable code, persistence, credential access, or hidden behavior.
Before installing, understand that this skill may be selected for broad browser or automation wording even when a more specific skill would be better. It appears safe as a planning/documentation helper, but users who manage many skills may want narrower trigger wording or explicit invocation only.
The trigger phrases are broad enough to match ordinary requests about productivity, browsers, automation, or practical workflows, which can cause the skill to activate outside its intended scope. In an agentic system, this increases the risk of prompt-routing mistakes, unintended tool use, and user confusion, especially when the skill may influence browser or automation behavior.
The manifest description contains very broad activation terms such as 'browser', 'automation', 'workflow', 'artifact', and 'analysis', which can cause the skill to match many unrelated user requests. Overbroad routing increases the chance this skill is invoked outside its intended scope, potentially overriding safer or more appropriate skills and causing unintended browser or automation guidance in sensitive contexts.
The trigger list includes generic keywords like 'browser', 'automation', 'cli', and 'bug fix' without qualifiers or exclusions, making accidental activation highly likely. In an agent ecosystem, this can lead to prompt-space hijacking by a loosely related skill, degrading least-privilege behavior and increasing the risk of inappropriate actions or unsafe advice being surfaced in the wrong context.
allow_implicit_invocation: true permits the skill to be invoked without an explicit user request, and this file provides no meaningful trigger constraints to limit when that happens. Combined with the broad productivity/browser framing, the skill could be activated unexpectedly and influence agent behavior in unrelated tasks, expanding attack surface for prompt injection or unsafe workflow execution.
The trigger sentences are extremely broad and partially consist of generic user phrasing like 'Help me', 'I need a practical workflow', and 'Use ... to handle ...', which can match many unrelated requests. In an agent-routing system, this can cause unintended invocation of the skill, misrouting user tasks, and unsafe overreach into contexts the skill was not meant to handle.
The trigger phrases are extremely generic and overlap with common user requests about browsers, automation, workflows, and productivity. This can cause the skill to activate outside its intended scope, injecting unintended instructions or taking over unrelated conversations, which is especially risky for an agent/browser-oriented skill that may influence automated actions.
In a zh-CN README, the primary behavioral instructions and invocation phrases are written in English, which can impose an implicit language requirement on users without opt-in. This may violate language/locale policy expectations because the file does not offer an alternative Chinese trigger set or clarify that English is optional.
The skill description defines activation conditions in very general terms, covering broad categories like work productivity, browser, headless, and automation without clear boundaries. In a multi-skill environment, this ambiguity increases the chance of accidental invocation and misrouting, which can degrade reliability and lead to unintended handling of user requests.
The trigger list is overly broad and includes generic terms like "browser", "automation", "cli", and "bug fix", which are common across many unrelated tasks. This can cause the skill to activate in contexts where it is not appropriate, potentially overriding more suitable skills or causing the agent to follow an irrelevant workflow.
The default_prompt contains a very broad invocation phrase and generic task wording such as work-productivity, browser, automation, checklist, and analysis support. This increases the chance the skill is triggered in unrelated conversations, causing prompt-scope confusion or unintended application of the skill in contexts the user did not explicitly request.
