Security audit
Security checks for vulnerabilities and agentic risk
This is a low-risk workflow helper with overly broad activation wording but no executable code, persistence, credential access, or destructive behavior.
This skill appears safe to install as a lightweight workflow helper. Review its broad trigger terms and implicit invocation setting if you want to avoid it being selected for unrelated web, API, or general productivity requests.
The skill description contains very broad invocation terms such as 'web', 'search', 'via', and generic requests for 'practical workflow, artifact, checklist, analysis, or implementation support'. This can cause the skill to activate for many unrelated user requests, leading to scope hijacking, incorrect tool selection, or accidental routing into a workflow the user did not intend.
The trigger keyword list includes highly ambiguous terms like 'web', 'search', 'via', 'api', 'user', and 'alternative', which are common in ordinary conversation. In an agent ecosystem, this makes unintended invocation likely, increasing the chance of prompt-routing mistakes, irrelevant skill takeover, and downstream unsafe behavior if the wrong workflow is applied to sensitive tasks.
The skill description states activation for broad categories such as work-productivity, tavily, web, search, and via, or for any request needing a workflow, checklist, analysis, or implementation support for the requirement. This is overly inclusive and can cause accidental routing of many normal productivity or web-search requests to this skill, creating prompt-scope confusion and unreliable behavior.
The trigger keywords are extremely broad and include common terms like "web", "search", "api", and "user", which can cause the skill to activate in many unrelated conversations. Over-broad activation increases the chance that the skill injects irrelevant workflow guidance into contexts where it was not intended, reducing reliability and potentially interfering with safer or more appropriate skills.
The trigger sentences are so broad and generic that they can match ordinary user requests unrelated to this specific Tavily-style workflow skill. That creates scope hijacking risk: the wrong skill may activate, causing unintended handling of user data, confusing behavior, or bypass of more appropriate, narrowly scoped skills.
The trigger phrases are broad, natural-language prompts that can match ordinary user requests unrelated to this skill’s intended scope. In an agent environment, this can cause unintended invocation, misrouting, or overshadowing of more appropriate skills, which increases the chance of unsafe or incorrect actions being taken under the wrong workflow.
The invocation guidance does not define clear boundaries for when the skill should or should not be used, and it mixes broad keywords like 'api', 'alternative', 'user', and 'bug fix' with generic trigger sentences. This makes accidental or excessive invocation more likely, which is especially risky for a skill tied to web/search workflows because it may prompt unnecessary external-query behavior or produce domain-mismatched assistance.
The trigger phrases are broad enough to match common, everyday requests such as asking for a practical workflow or help with bug fixing, which can cause the skill to activate outside its intended Tavily/web-workflow scope. In an agent ecosystem, overbroad activation can route unrelated user tasks into this skill, increasing the chance of inappropriate tool use, misleading outputs, or bypass of more suitable safeguards in specialized skills.
The example trigger sentences use broad phrasing like 'Help me' and 'I need a practical workflow' without clearly delimiting when the skill should or should not activate. These examples reinforce overbroad matching behavior and make accidental invocation more likely, though they are somewhat less dangerous than the keyword list itself because they are illustrative rather than direct matching logic.
该文件整体为 zh-CN 中文版本，但文档中未说明这是可选本地化输出，也未向用户提供语言选择。按照语言/地区策略，若技能默认强制特定语言而无用户选择，可能构成自然语言策略违规。
The default prompt uses broad, natural-language trigger terms tied to common tasks ('work-productivity', 'web', 'search', 'workflow'), which increases the chance of unintended or overly frequent invocation. Because implicit invocation is enabled, this can cause the skill to activate outside narrowly intended contexts, injecting its behavior or instructions into unrelated user workflows.
