Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with broad activation wording but no code execution, persistence, credential access, or hidden data handling.
Install only if you want an agent response helper for ontology-style or typed workflow planning. Be aware it may be selected for some generic productivity, workflow, knowledge-graph, or bug-fix prompts because its trigger wording is broad; disable implicit invocation or narrow triggers if precise routing matters.
The skill description and use conditions are broad enough to match many ordinary productivity or knowledge-work requests, which can cause unintended invocation. Over-broad activation increases the chance that this skill intercepts tasks outside its intended scope, leading to confused routing, reduced safety review quality, or inappropriate guidance in contexts that need a more specialized skill.
The keyword list includes highly generic terms such as 'typed', 'knowledge', 'graph', 'structured', 'memory', and 'creating', which are common across many unrelated requests. This makes accidental activation likely and can let the skill override more appropriate tools or workflows, especially in automated routing systems that depend on keyword matching.
The trigger sentence is broad enough to match ordinary user phrasing rather than a narrowly scoped skill invocation. In agent ecosystems, overly generic activation text can cause unintended skill selection, leading to prompt-routing errors, irrelevant workflow execution, or accidental handling of requests outside the skill's intended boundary.
The trigger examples are generic enough that normal user requests about productivity, workflows, bug fixing, or hardening could unintentionally activate this skill outside the author’s intended scope. In an agent ecosystem, unintended invocation can cause the wrong workflow to run, leading to irrelevant actions, confused task routing, or application of ontology/typed-process guidance to requests that did not explicitly ask for it.
The trigger phrases are broad, generic, and partially templated around common requests like asking for help or a practical workflow, which increases the chance the skill is invoked when the user did not explicitly intend it. In an agent ecosystem, unintended invocation can route user data or task flow into the wrong skill, causing confusion, mis-execution, or unsafe automation chaining.
The example triggers are written as ordinary natural-language requests and do not establish a clear invocation boundary, so similar everyday user phrasing may activate the skill unintentionally. In a multi-skill environment, underspecified examples reinforce ambiguous routing behavior and make it harder to ensure the right capability handles the request.
用途描述写成了宽泛的主题词加上“需要实用流程、产物、检查清单、分析或实现支持时使用”，边界非常模糊，几乎可覆盖许多普通生产力或知识整理请求。这会使调度器难以区分该技能与其他相邻技能的职责，增加错误路由、权限外使用或输出不符合用户真实任务的概率。
触发关键词包含 knowledge、graph、structured、creating、bug fix 等高频且跨领域词汇，容易在大量无关对话中误触发该技能。误触发会让代理在错误上下文中调用不相关工作流，导致结果偏题、覆盖更合适的技能选择，降低可靠性并可能把本应受限的操作扩展到不必要的场景。
The default prompt is extremely broad and includes generic terms like work-productivity, workflow, checklist, analysis, and implementation support, which can match many ordinary user requests unrelated to this specific skill. This increases the chance of unintentional routing or invocation, causing the skill to influence conversations outside its intended scope and potentially exposing users to unintended instructions or side effects.
Enabling implicit invocation without strong trigger constraints allows the platform to auto-select this skill based on weak semantic matches. Because the skill description and prompt are broad, implicit invocation compounds the risk of accidental activation, making this materially more dangerous in context than either setting alone.
The invocation guidance mixes ambiguous natural-language examples with broad keywords such as work-productivity, knowledge, and structured, which are common across many unrelated tasks. This increases the risk of unintentional invocation or skill overreach, especially in multi-skill agents where routing logic may use fuzzy matching and select this skill for requests it should not handle.
The file is presented as a Chinese README, but the actual usage instructions and trigger phrases are only in English. This can amount to forcing a specific language for core operation without explicitly offering the user a language or locale choice.
