Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with broad trigger wording, but it does not request sensitive access, persistence, credential use, code execution, or hidden behavior.
Installers should be aware that this skill may activate on broad productivity, knowledge, graph, or bug-fix wording. Review whether those triggers fit your routing preferences, but the package itself does not show hidden execution, credential access, persistence, or destructive behavior.
The skill description and usage guidance are broad enough to match many ordinary productivity or implementation requests, which can cause the skill to activate outside its intended scope. Over-broad activation increases the chance that an irrelevant or lower-safety workflow is injected into conversations, reducing reliability and potentially bypassing more appropriate specialized skills.
The keyword list includes ambiguous terms such as "typed," "knowledge," "graph," "structured," "memory," and "creating," which are common across many unrelated tasks. This makes accidental invocation highly likely, and in agent systems that auto-select skills, such ambiguity can systematically route users into the wrong skill and produce unsafe or misleading assistance.
The trigger phrases are generic enough to match normal user requests about productivity, ontology workflows, bug fixing, and hardening, which increases the chance this skill is invoked unintentionally. In an agent ecosystem, over-broad activation can hijack routing, cause the wrong skill to handle sensitive tasks, and amplify any unsafe instructions present in downstream skill files.
The trigger phrases are broad enough to match many ordinary user requests about productivity, workflows, knowledge, graphs, or bug fixing, which can cause the skill to activate outside its intended scope. In an agent ecosystem, overbroad routing is dangerous because it can hijack unrelated tasks, produce irrelevant actions, and increase the chance of unsafe downstream behavior by invoking the wrong capability.
The example trigger sentences are effectively generic request wrappers around a broad marketing-style requirement statement, without defining boundaries for when the skill should or should not be used. This trains invocation behavior toward vague matching, which can expand the skill's activation surface and degrade routing precision.
The activation criteria in the description are phrased too broadly, covering users who ask for practical workflows, artifacts, checklists, analysis, or implementation support across several generic concepts. In a skill-selection system, this overbreadth increases the chance of accidental invocation and misapplication of the skill to unrelated tasks.
The trigger keywords are very broad and include generic terms like “knowledge”, “graph”, “structured”, and “bug fix”, which can overlap with many unrelated requests. This can cause the skill to activate outside its intended domain, leading to incorrect routing, unexpected behavior, or overshadowing of more appropriate skills.
The default_prompt is vague and broadly phrased, and the skill also allows implicit invocation. Together, this increases the chance that the agent may trigger this skill in situations only loosely related to the user's intent, causing unintended workflow execution or irrelevant guidance. In a skill intended to provide practical implementation support across broad topics, ambiguous activation criteria can lead to prompt-routing mistakes and reduce safety by invoking the wrong capability at the wrong time.
The trigger sentences are broad enough to match common user requests about workflows, bug fixes, hardening, or practical help, which can cause the skill to activate outside its intended scope. In an agent-routing context, this creates prompt/skill selection confusion and can lead to inappropriate invocation, unexpected data exposure to the skill, or reduced reliability of downstream task handling.
A locale-specific README for zh-CN includes substantial English-only instructional content such as the title, demand description, and trigger phrases. This can amount to an implicit language requirement without user opt-in, and no explanation is given for why the Chinese-language document requires English interaction.
