Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with no executable code or persistence, though its activation terms are too broad and may trigger unexpectedly.
This skill is reasonable to install for Tavily or web-search workflow support, but users should be aware it may activate on generic web/search/API requests unless the platform or publisher narrows the triggers or disables implicit invocation.
The skill description and activation scope are extremely broad, covering common concepts like work productivity, web, search, and implementation support. This can cause the skill to be invoked for many unrelated requests, increasing the chance that users receive mismatched instructions or that the skill overrides a more appropriate, narrower skill.
The trigger list includes highly ambiguous terms such as 'web', 'search', 'api', 'user', and 'alternative', which are common across many unrelated conversations. In an agent-routing context, this can lead to frequent unintended invocation, causing prompt hijacking of normal workflows and reducing reliability or safety by inserting irrelevant behavior into tasks.
The invocation guidance is ambiguous because it mixes broad keywords such as 'web', 'search', 'via', and 'user' with loosely defined trigger sentences, leaving the activation boundary unclear. This makes accidental invocation materially more likely, especially in productivity or web-research conversations where these terms are common, and can lead to unintended execution of the skill in unrelated contexts.
The trigger examples are written as natural-language requests with broad, everyday phrasing, which can cause the skill to activate when a user is merely discussing Tavily-style workflows rather than intentionally invoking this specific skill. In an agent ecosystem, overly permissive trigger text increases the risk of unintended routing, causing the wrong skill to run and potentially exposing user data or producing actions outside the user's intent.
The trigger phrases are broad, generic, and include common terms like 'user', 'api', 'web', and 'bug fix', which can cause the skill to activate for many ordinary requests outside its intended scope. In an agent environment, this increases the chance of incorrect routing, accidental invocation, and unintended influence over unrelated tasks, reducing reliability and potentially exposing users to workflows they did not request.
The example trigger phrases are generic and largely restate the broad requirement text without clear boundaries for valid invocation. This makes it difficult for maintainers or routing systems to distinguish intended from unintended use, increasing accidental activation and the risk of low-quality or irrelevant task handling.
The skill description defines activation conditions with vague language such as broad topical references and 'or needs practical workflow/artifact/checklist/analysis/implementation support', which leaves the activation boundary unclear. Ambiguous routing criteria can cause this skill to be selected for unrelated requests, potentially exposing users to irrelevant instructions or overriding more appropriate skills.
The trigger keyword list is overly broad and includes generic terms like 'web', 'user', 'api', and 'search', which can cause the skill to activate in many unrelated contexts. In an agent environment, this increases the chance of prompt/context hijacking at routing time, causing the wrong skill to influence tasks it was not intended to handle.
The default prompt uses a very broad invocation phrase tied to generic terms like work-productivity, tavily, web, search, workflow, checklist, analysis, or implementation support, which can cause the skill to trigger in many unrelated contexts. Combined with allow_implicit_invocation: true, this increases the chance of unintended activation, prompt-scope capture, and accidental routing of user requests into this skill when the user did not clearly ask for it.
The trigger sentence is overly broad and resembles normal user phrasing, which can cause the skill to activate when the user did not specifically intend to invoke it. In an agent environment, unintended invocation can route tasks through the wrong workflow, causing confusion, irrelevant actions, or unsafe tool usage if downstream web/search behavior is enabled.
This trigger uses a generic request pattern that does not clearly bound activation to a narrow use case, increasing the chance of accidental matching across ordinary productivity requests. Because this skill is framed broadly around workflows, artifacts, analysis, and implementation help, ambiguous activation expands its reach and makes misrouting more likely.
The document is labeled as a Chinese README, but substantial instructional content and trigger phrases remain in English. This can impose a language expectation without explicit user opt-in or a documented language-selection policy.
The description includes the Chinese term "搜索" embedded in otherwise English instructions, which imposes a mixed-language presentation without stating that the user can choose language or locale. This can conflict with language/locale policy expectations when no opt-in is provided.
该技能文件整体以中文撰写，文件名也明确为 zh-CN，但正文未说明这是可选本地化版本，也未给出用户可选择其他语言的提示。若此内容被作为技能默认说明使用，可能构成未获用户选择的语言/地区约束。
The phrase "Tavily 搜索-style workflows" mixes English and Chinese, which may impose a language or locale expectation without documenting opt-in or region-specific justification. Under the policy, language constraints should either be optional for the user or clearly explained.
